Evolution and Innovation: A Comparative Genomic Analysis of NBS Domain Genes in Plant Immunity

Abstract

This article provides a comprehensive synthesis of current research on Nucleotide-Binding Site (NBS) domain genes, the primary class of disease resistance (R) genes in plants. Covering foundational concepts to advanced applications, we explore the remarkable diversity and evolution of NBS genes across land plants, from mosses to major crops. The review details state-of-the-art bioinformatics and machine learning methodologies for gene identification, addresses key challenges in analyzing this complex gene family, and presents case studies on the functional validation of specific NBS genes against viral, fungal, and bacterial pathogens. Tailored for researchers and scientists in plant genomics and drug development, this analysis highlights how understanding plant immune receptors can inform broader strategies for disease resistance and therapeutic discovery.

Unveiling the Diversity and Evolutionary History of NBS Domain Genes

Plant immunity relies on a sophisticated innate immune system where nucleotide-binding site (NBS) domain genes play a pivotal role as the largest class of plant disease resistance (R) genes [1]. These genes encode proteins that are vital for plant defense, enabling the detection of pathogen-derived molecules and initiating robust defense responses [2]. The NBS domain forms the core of a larger superfamily of proteins known as NLRs (Nucleotide-binding Leucine-rich Repeat receptors) [3] [4]. These intracellular immune receptors are modular proteins, typically consisting of a variable N-terminal domain, a central NBS (NB-ARC) domain, and C-terminal leucine-rich repeats (LRRs) [1] [2]. The NBS domain functions as a molecular switch, binding and hydrolyzing ATP/GTP to provide energy for downstream signaling processes [5] [6], while the LRR domain is primarily involved in pathogen recognition [7] [5].

This guide provides a comparative analysis of NBS domain genes across plant species, detailing their classification, distribution, and evolution. We present standardized experimental protocols for their identification and characterization, supported by quantitative data and visualizations of immune signaling pathways, to serve as a resource for researchers and drug development professionals.

Classification and Genomic Distribution of NBS Genes

Classification of NBS Gene Subfamilies

Based on the structure of the N-terminal domain, NBS-encoding genes are classified into distinct subfamilies, which have diverged to perform specialized functions in plant immunity [3] [1].

• TNL (TIR-NBS-LRR): Characterized by a Toll/Interleukin-1 Receptor (TIR) domain. These genes are predominantly found in dicots and are absent in most monocots [7] [8].
• CNL (CC-NBS-LRR): Feature a Coiled-Coil (CC) domain at the N-terminus. This subclass is present in both monocots and dicots and often represents the most abundant type [1] [9].
• RNL (RPW8-NBS-LRR): Contain a Resistance to Powdery Mildew 8 (RPW8) domain. Unlike TNLs and CNLs, RNLs typically function downstream in signal transduction rather than in direct pathogen recognition [1] [9].

In addition to these canonical architectures, plants possess numerous truncated forms (lacking LRRs or N-terminal domains) and NLRs with Integrated Domains (NLR-IDs). These integrated domains can act as "baits" for pathogen effectors, enabling novel recognition capabilities [4].

Comparative Quantitative Analysis Across Species

The number and composition of NBS genes vary dramatically across plant species, influenced by their evolutionary history and pathogen pressures. The table below summarizes a comparative analysis from recent studies.

Table 1: Comparative Analysis of NBS-Encoding Genes Across Plant Species

Plant Species	Family	Total NBS Genes	CNL	TNL	RNL	Notable Features	Citation
Arabidopsis thaliana	Brassicaceae	167	69 (41%)	92 (55%)	6 (4%)	Model dicot with balanced TNL/CNL	[8]
Brassica oleracea	Brassicaceae	157	89 (57%)	62 (39%)	6 (4%)	CNL expansion post-WGT	[8]
Xanthoceras sorbifolium	Sapindaceae	180	155 (86%)	23 (13%)	2 (1%)	"First expansion then contraction" pattern	[1]
Dinnocarpus longan	Sapindaceae	568	502 (88%)	43 (8%)	23 (4%)	Strong recent gene expansion	[1]
Vernicia montana	Euphorbiaceae	149	98 (66%)	12 (8%)	2 (1%)	Resistant to Fusarium wilt	[5]
Vernicia fordii	Euphorbiaceae	90	49 (54%)	0 (0%)	0 (0%)	Susceptible to Fusarium wilt; TNL loss	[5]
Akebia trifoliata	Lardizabalaceae	73	50 (68%)	19 (26%)	4 (5%)	Low number; uneven chromosomal distribution	[6]
Dendrobium officinale	Orchidaceae	74	10 (14%)	0 (0%)	N/A	No TNL genes identified; common in monocots	[7]

The data reveals several key trends. First, the number of NBS genes is highly dynamic, even within the same family, as seen in the Sapindaceae species where D. longan has over three times the number of genes found in X. sorbifolium [1]. Second, the dominance of the CNL subclass is a recurring theme across many angiosperms [1] [9]. Third, the absence of TNLs in monocots like orchids is a well-established phenomenon, potentially driven by the deficiency of the NRG1/SAG101 signaling pathway [7]. Finally, comparative analyses of resistant and susceptible varieties, such as in tung trees (Vernicia), can pinpoint specific gene losses (e.g., TNLs in susceptible V. fordii) associated with disease susceptibility [5].

Genomic Architecture and Evolution

NBS-encoding genes are not randomly distributed within plant genomes. They are frequently organized in clusters located in hot-spot regions on chromosomes [2] [6]. These clusters can be homogeneous (containing the same type of NLR) or heterogeneous (containing diverse NLR classes or even mixed with other receptor genes) [2]. This arrangement is primarily driven by gene duplication events, including tandem duplications and whole-genome duplications (WGD), which facilitate the birth of new resistance specificities [2] [8]. Following duplication, genes undergo a process of birth and death, with some copies being preserved through natural selection while others are lost or become pseudogenes [2]. This dynamic leads to the distinct evolutionary patterns observed in different plant lineages, such as "expansion and contraction" or "continuous expansion" [1] [9].

Research into NBS domain genes relies on a suite of bioinformatic tools and genomic resources. The following table outlines key solutions for identification and characterization.

Table 2: Essential Research Reagents and Resources for NBS Gene Analysis

Research Tool	Type	Primary Function in NBS Research	Example Usage
HMMER	Software	Identifying NBS domain-containing proteins in genome assemblies using hidden Markov models.	Search with NB-ARC (PF00931) HMM profile [3] [1] [5].
Pfam / NCBI-CDD	Database	Validating the presence of protein domains (NBS, TIR, CC, LRR, RPW8).	Confirm domain architecture of candidate genes [1] [7] [6].
OrthoFinder	Software	Inferring orthogroups and gene families across multiple species.	Reconstructing evolutionary history and classifying NBS genes [3].
MEME Suite	Web Tool	Discovering conserved protein motifs within NBS domains and other regions.	Identifying structural motifs specific to CNL, TNL, or RNL subfamilies [9] [6].
RNA-seq Data	Data	Profiling gene expression under various conditions (biotic/abiotic stress, different tissues).	Identifying differentially expressed NBS genes in resistant vs. susceptible cultivars [3] [5].
Virus-Induced Gene Silencing (VIGS)	Experimental Method	Functional validation of NBS genes through transient silencing.	Knocking down a candidate NBS gene (e.g., GaNBS) to test its role in disease resistance [3] [5].

Experimental Protocols for Identification and Functional Characterization

A standardized pipeline for genome-wide identification and functional analysis of NBS genes is critical for comparative studies. The workflow below outlines the key stages from identification to functional validation.

Diagram 1: Experimental workflow for NBS gene analysis

Detailed Methodological Protocols

Protocol 1: Genome-Wide Identification of NBS-Encoding Genes

This protocol is adapted from methodologies used in multiple comparative genomic studies [3] [1] [8].

• Data Collection: Obtain the latest genome assembly and annotation files for the target species from public databases (e.g., NCBI, Phytozome, species-specific databases).
• HMMER Search: Perform a hidden Markov model (HMM) search against the predicted proteome using the NB-ARC domain model (Pfam accession: PF00931). Use the hmmsearch tool from the HMMER package with default parameters or a stringent e-value cutoff (e.g., 1.1e-50) [3].
• BLAST Search: Conduct a complementary BLASTP search using known NBS protein sequences as queries against the proteome, with an e-value threshold of 1.0 [1] [9].
• Data Curation: Merge the candidate sequences from both searches and remove redundant entries.
• Domain Validation: Confirm the presence of the NBS domain in all remaining candidates by analyzing them against the Pfam and NCBI Conserved Domain Database (CDD) with an e-value of 10⁻⁴ [1] [7]. Identify additional domains (TIR, CC, LRR, RPW8) using HMM profiles and specialized tools like COILS or Marcoil for CC domains [8].

Protocol 2: Functional Validation via Virus-Induced Gene Silencing (VIGS)

VIGS is a powerful technique for rapid functional characterization, as demonstrated in studies on cotton and tung tree NBS genes [3] [5].

• Candidate Gene Selection: Select a target NBS gene (e.g., a gene highly expressed in resistant cultivars upon pathogen infection).
• Vector Construction: Clone a 200-300 bp fragment specific to the target gene into a VIGS vector (e.g., TRV-based pYL156 or pYL279).
• Plant Material & Inoculation: Grow plants (e.g., resistant cotton or tung tree) to the 2-4 leaf stage. Inoculate by agroinfiltration, where Agrobacterium tumefaciens harboring the VIGS construct is injected into the leaves.
• Control Groups: Include plants inoculated with an empty vector (negative control) and a vector carrying a phytoene desaturase (PDS) gene fragment (positive control for silencing efficiency).
• Silencing Confirmation: After 2-3 weeks, assess silencing efficiency by measuring target gene transcript levels in silenced plants compared to controls using quantitative RT-PCR (qRT-PCR).
• Phenotypic Assay: Challenge the silenced and control plants with the target pathogen. Monitor disease symptoms, measure pathogen biomass, and assess physiological changes to determine the role of the silenced NBS gene in disease resistance.

NBS-Mediated Immune Signaling Pathways

NBS-LRR proteins are central to effector-triggered immunity (ETI), a robust immune response that often culminates in a hypersensitive response (HR) to restrict pathogen spread [7] [2]. The signaling pathways differ based on the NBS subfamily involved. The diagram below illustrates the core ETI signaling pathway and the distinct roles of TNL and CNL receptors.

Diagram 2: Core ETI signaling pathways

As depicted, TNL and CNL proteins act as sensors that directly or indirectly recognize pathogen effectors [1] [9]. This recognition triggers a conformational change in the NBS domain, facilitating nucleotide exchange (ADP to ATP) and activating the receptor [2]. Activated TNLs signal through the EDS1-PAD4-SAG101 protein complex, while the signaling pathway for CNLs is less defined but may involve other components [7]. Both pathways converge on RNL proteins (e.g., NRG1, ADR1), which function as helper NLRs to transduce the defense signal downstream [1] [9]. This leads to the activation of defense genes, a burst of reactive oxygen species, and often the initiation of the hypersensitive response, a form of programmed cell death at the infection site [2].

Plant immunity relies on a sophisticated surveillance system mediated by intracellular receptors known as nucleotide-binding leucine-rich repeat receptors (NLRs). These proteins detect pathogen effector molecules and initiate robust defense responses, culminating in effector-triggered immunity (ETI) [10] [11]. NLRs are classified into major structural classes based on their N-terminal domains, which dictate their signaling mechanisms and functional specializations. This guide provides a comparative analysis of the Toll/Interleukin-1 Receptor (TNL), Coiled-Coil (CNL), and Non-TNL classes, examining their structural features, evolutionary patterns, and activation mechanisms to inform research and development in plant disease resistance.

Classification and Structural Features

The NLR superfamily in plants is divided into two major classes based on the N-terminal domain, with a third category encompassing atypical configurations.

Table 1: Major Structural Classes of Plant NLR Genes

Class	N-Terminal Domain	Key Domains & Architecture	Prevalence & Distribution	Representative Examples
TNL	Toll/Interleukin-1 Receptor (TIR)	TIR-NBS-LRR; TIR domain has enzymatic activity (NAD+ cleavage)	Abundant in dicots; scarce in most monocots [11]	MRT1, MRT2, MIST1 (Arabidopsis) [10]; RPP1, RPS4 [11]
CNL	Coiled-Coil (CC)	CC-NBS-LRR; CC domain forms signaling-competent complexes	Most abundant class in monocots; found across all angiosperms [3] [11]	Sr33, MLA10, Rx, RPS2, RPS5 (Arabidopsis and cereals) [11]; ZAR1 (forms resistosome) [10]
Non-TNL / nTNL	Non-TIR, various domains	Includes RPW8-NLRs (RNLs); other atypical domain architectures	Least abundant class; RNLs often function as "helper NLRs" [10] [3]	ADR1, NRG1 (helper RNLs) [3]; Proteins with TIR-NBS-TIR-Cupin1, Sugartr-NBS domains [3]

Table 2: Functional and Evolutionary Characteristics

Characteristic	TNLs	CNLs	Non-TNLs / nTNLs
Primary Signaling Mechanism	TIR domain forms holoenzyme, produces signaling molecules (e.g., cADPR) [10]	CC domain inserts into plasma membrane, potential ion channel activity [10] [11]	Varied; RNLs signal through CC domain and can be required for TNL/CNL immunity [10] [3]
Activation Complex	TIR-domain tetramer [10]	CC-domain pentamer (resistosome) [10]	Not fully characterized for all types
Regulatory Mechanisms	Targeted by miRNAs (e.g., miR825-5p); generate phasiRNAs for amplified silencing [10]	Regulated by intramolecular interactions (e.g., EDVID motif with NB domain) [11]	Less studied; likely subject to transcriptional and post-transcriptional control
Evolutionary Dynamics	Ancient origin; expanded in dicots; birth-and-death evolution with gene loss/duplication [3] [12]	Massive expansion in flowering plants; high sequence diversity in CC domain [3] [11]	Includes conserved helper NLRs (RNLs) and lineage-specific genes with novel domain fusions [3]

Experimental Approaches for Comparative Analysis

Genome-Wide Identification and Classification

Protocol for NBS Domain Identification and Classification:

• Sequence Retrieval: Obtain genome assemblies and protein sequences from public databases (e.g., NCBI, Phytozome) [3].
• Domain Screening: Use HMMER-based tools (e.g., PfamScan.pl) with the NB-ARC (PF00931) Hidden Markov Model (HMM) profile to identify candidate NBS-containing genes. A typical e-value cutoff is 1.1e-50 for stringency [3].
• Architecture Classification: Analyze the domain architecture of identified genes using HMM scans or InterProScan to detect associated N-terminal (TIR, CC, RPW8) and C-terminal (LRR) domains. Classify genes into TNL, CNL, or Non-TNL categories based on the presence of TIR, CC, or other N-terminal domains, respectively [3].
• Orthogroup Analysis: Use tools like OrthoFinder with the DIAMOND algorithm for sequence similarity searches and the MCL algorithm for clustering to define orthogroups (groups of genes descended from a single gene in the last common ancestor). This helps identify core, conserved lineages and species-specific expansions [3].

Functional Validation through Gene Silencing

Virus-Induced Gene Silencing (VIGS) Protocol:

• Candidate Gene Selection: Select a target NBS gene (e.g., GaNBS from cotton) based on expression profiles under stress [3].
• Vector Construction: Clone a ~300-500 bp fragment of the target gene into a VIGS vector (e.g., TRV-based pYL192 series).
• Plant Inoculation: Agrobacterium tumefaciens strains harboring the VIGS construct are grown, resuspended in induction media, and infiltrated into cotyledons or true leaves of young plants.
• Phenotyping: After silencing is established (typically 2-3 weeks post-infiltration), challenge plants with the target pathogen. Assess disease symptoms, pathogen biomass, and compare to control plants (silenced with an empty vector) [3].
• Validation: Use qRT-PCR to confirm the reduction of target gene transcript levels and quantify pathogen titer.

Diagram Title: NLR Class Signaling Pathways

The Scientist's Toolkit: Key Research Reagents and Solutions

Table 3: Essential Research Reagents for NLR Gene Analysis

Reagent / Resource	Function / Application	Example Use Case
HMMER/PfamScan	Identifies NBS (NB-ARC) domains and other architectural domains in protein sequences [3].	Genome-wide identification and classification of NLRs into TNL, CNL, and Non-TNL classes [3].
OrthoFinder	Infers orthogroups and gene families from whole-genome data; elucidates evolutionary relationships [3].	Comparing NLR repertoires across species to identify conserved orthogroups (e.g., OG2, OG6) and lineage-specific expansions [3].
VIGS Vectors (e.g., TRV-based)	Enables transient, sequence-specific silencing of target genes in plants [3].	Functional validation of candidate NBS genes (e.g., GaNBS in cotton) by assessing susceptibility upon silencing [3].
NBS Profiling Primers	Degenerate primers targeting conserved NBS motifs (P-loop, Kinase-2, GLPL) amplify NBS tags for sequencing [13].	Profiling the R-gene repertoire and allelic diversity in plant populations without whole-genome sequencing [13].
RNA-seq Datasets (e.g., from IPF, CottonFGD)	Provides expression data (FPKM) across tissues and stress conditions [3].	Identifying NLR genes with putative defense roles based on upregulation under biotic/abiotic stress [3].
Methyl 6-fluorohexanoate	Methyl 6-Fluorohexanoate|CAS 333-07-3|Research Chemical	High-purity Methyl 6-Fluorohexanoate (CAS 333-07-3), a key synthon for PET tracer and bioconjugation research. For Research Use Only. Not for human or veterinary use.
Ikarisoside-F	Ikarisoside-F, MF:C31H36O14, MW:632.6 g/mol	Chemical Reagent

The structural dichotomy between TNLs and CNLs represents a fundamental evolutionary strategy in plant immunity, with distinct signaling mechanisms (TIR-enzymatic activity versus CC-mediated resistosome formation) converging on effective pathogen resistance [10] [11]. Non-TNLs, particularly RNLs, play critical, complementary roles as helper NLRs. Comparative genomics reveals that these gene families undergo dynamic evolution, driven by birth-and-death processes and gene duplication, resulting in lineage-specific expansions and losses [3] [12]. A deep understanding of these classes, their interactions, and regulatory networks—including miRNA-mediated silencing as exemplified by the miR825-5p/TNL module [10]—provides a robust foundation for developing durable, broad-spectrum resistance in crops through modern biotechnological approaches.

The evolutionary transition of plants from aquatic to terrestrial environments necessitated the development of sophisticated immune mechanisms to combat emerging pathogens. Central to this adaptive innovation are nucleotide-binding site (NBS) domain genes, which encode one of the largest and most critical families of plant disease resistance (R) genes [3] [14]. These genes provide plants with the capacity to recognize diverse pathogens and initiate robust defense responses [14].

The NBS-leucine-rich repeat (LRR) gene family exhibits remarkable structural diversity and evolutionary dynamics across the plant kingdom [15]. While comprehensive surveys have documented their expansion in numerous angiosperm species [16], studies in early land plants like bryophytes have revealed unexpected diversity and novel structural classes [16] [17]. Recent super-pangenome analyses of 123 bryophyte genomes further demonstrate that these non-vascular plants possess a substantially larger gene family space than vascular plants, with numerous unique and lineage-specific gene families [18].

This comparative guide objectively analyzes the diversification of NBS domain genes from bryophytes to angiosperms, synthesizing experimental data to elucidate evolutionary patterns and functional conservation. We provide detailed methodologies for key experiments and visualization of signaling pathways to support research in plant immunity and drug development.

Evolutionary History and Diversification Patterns

Deep Evolutionary Origins

NBS-LRR genes originated during early plant colonization of land, with the NBS domain combining with LRR domains coinciding with this evolutionary transition [16] [17]. Investigations across diverse plant lineages indicate that the common ancestor of bryophytes and vascular plants possessed the genetic machinery for NBS-mediated immunity, though the specific domain architectures have undergone substantial lineage-specific evolution [16] [18].

Bryophytes, as the sister group to all vascular plants, provide critical insights into the early evolution of plant immune genes. A comprehensive analysis of 138 bryophyte genomes revealed they possess a cumulative 637,597 nonredundant gene families compared to 373,581 in vascular plants, despite bryophytes having fewer genes per genome on average [18]. This expanded gene family diversity includes numerous NBS domain genes with unique domain architectures not observed in higher plants [16] [17].

Lineage-Specific Expansions and Losses

Analyses across angiosperms reveal dynamic expansion patterns influenced by both whole-genome duplications and small-scale duplication events [3] [15]. The three anciently diverged NBS-LRR classes (TNLs, CNLs, and RNLs) expanded into at least 23 lineages in the common ancestor of angiosperms [15]. A pattern of gradual expansion during the first 100 million years of angiosperm evolution was observed for CNLs, while TNL numbers remained relatively stable during this period [15].

Notably, an intense expansion of both TNL and CNL genes commenced at the Cretaceous-Paleogene boundary, potentially reflecting convergent adaptive responses to dramatic environmental changes and increased fungal diversity during this period [15]. Lineage-specific losses also occurred, with TNL genes completely absent from monocot genomes despite their presence in basal angiosperms like Amborella trichopoda [15].

Table 1: Comparative Analysis of NBS Domain Genes Across Plant Lineages

Plant Group	Representative Species	Total NBS Genes	TNLs	CNLs	RNLs	Unique Features
Bryophytes	Physcomitrella patens	65	9	11	-	PK-NBS-LRR (PNL) class [16]
Bryophytes	Marchantia polymorpha	43	-	7	-	Hydrolase-NBS-LRR (HNL) class [16]
Basal Angiosperm	Amborella trichopoda	105	15	89	1	Represents ancestral angiosperm NBS repertoire [15]
Eudicots	Arabidopsis thaliana	~150	~62	~88	-	Well-characterized reference [14]
Eudicots	Medicago truncatula	571	-	-	-	Highest number among surveyed angiosperms [15]
Monocots	Oryza sativa	>400	0	>400	-	Complete absence of TNL class [15]

Structural and Functional Classification

Domain Architecture Diversity

NBS domain genes typically exhibit a modular structure with an N-terminal domain, central NBS domain, and C-terminal LRR region [14]. Based on N-terminal domain identity, these genes are primarily classified into TIR-NBS-LRR (TNL) and coiled-coil-NBS-LRR (CNL) classes [14] [15]. A third class, RPW8-NBS-LRR (RNL), functions as scaffold proteins in defense signaling [15].

Beyond these canonical classes, bryophytes possess novel structural variants not found in angiosperms. In the moss Physcomitrella patens, researchers identified a PK-NBS-LRR (PNL) class characterized by an N-terminal protein kinase domain [16] [17]. Liverworts like Marchantia polymorpha possess a distinct Hydrolase-NBS-LRR (HNL) class featuring an N-terminal α/β-hydrolase domain [16] [17]. These novel classes exhibit unique intron positions and phase characteristics, suggesting independent evolutionary origins [16].

Motif Conservation and Variation

The NBS domain contains several conserved motifs (P-loop, RNBS-A, Kinase-2, RNBS-B, RNBS-C, GLPL, RNBS-D, and MHDV) that facilitate nucleotide binding and molecular switch functions [16] [14]. Phylogenetic analyses reveal closer relationships between HNL, PNL, and TNL classes, with CNLs representing a more divergent lineage [16].

Table 2: Conserved Motifs in the NBS Domain and Their Functional Roles

Motif Name	Consensus Sequence	Position	Functional Role	Conservation Across Plant Lineages
P-loop	GxPGSGKS	N-terminus	ATP/GTP binding	Universal in all plant NBS domains [16]
RNBS-A	FLHIACxF	After P-loop	Domain stability	Divergent between TNL/CNL classes [16]
Kinase-2	LVLDDVW	Middle	ATP hydrolysis	Highly conserved [16]
RNBS-B	GLPLAL	Middle	Domain folding	Variable [16]
RNBS-C	GSRIIITTRD	Middle	Unknown	Divergent between TNL/CNL classes [16]
GLPL	GLPLA	C-terminus	LRR interaction	Highly conserved [16]
RNBS-D	CFAL	C-terminus	Signaling regulation	Divergent between TNL/CNL classes [16]
MHDV	MHDIV	C-terminus	Nucleotide exchange	Highly conserved [16]

Genomic Distribution and Organization

Chromosomal Arrangement and Gene Clustering

NBS-encoding genes typically display non-random chromosomal distribution, frequently organized in clusters resulting from both segmental and tandem duplication events [19] [14]. Comparative analyses in asparagus species revealed that NLR genes in A. officinalis, A. kiusianus, and A. setaceus all exhibit clustering patterns across chromosomes, with adjacent NLR pairs often separated by ≤8 genes [19].

This clustering facilitates unequal crossing-over and gene conversion, generating variation in copy number and sequence diversity [14]. Studies in lettuce have identified two evolutionary patterns: type I genes evolve rapidly with frequent gene conversions, while type II genes evolve slowly with rare gene conversion events [14].

Evolutionary Dynamics and Selection Pressures

NBS genes evolve through a birth-and-death process characterized by gene duplication, sequence diversification, and pseudogenization [14] [20]. Different protein domains experience distinct selection pressures, with the NBS domain typically under purifying selection to maintain functional integrity, while LRR regions experience diversifying selection to generate recognition specificities [14].

Recent research has revealed that microRNA targeting represents an important regulatory mechanism for NBS-LRR genes, with diverse miRNA families emerging to target highly duplicated NBS-LRRs [21]. These miRNA-NBS-LRR interactions likely help balance the benefits and costs of maintaining large NBS-LRR repertoires [21].

Experimental Approaches and Methodologies

Genome-Wide Identification and Classification

Protocol 1: Identification of NBS Domain Genes

• Data Collection: Obtain genome assemblies and annotation files from public databases (NCBI, Phytozome, Plaza) [3].
• Domain Screening: Use PfamScan.pl HMM search script with default e-value (1.1e-50) and Pfam-A_hmm model to identify genes containing NB-ARC domains [3].
• Sequence Validation: Extract candidate sequences and validate through domain architecture analysis using InterProScan and NCBI's Batch CD-Search [19].
• Classification: Categorize genes based on complete domain architecture using Pfam and PRGdb databases [19].

Protocol 2: Evolutionary and Phylogenetic Analysis

• Orthogroup Construction: Use OrthoFinder v2.5.1 with DIAMOND for sequence similarity searches and MCL clustering algorithm for gene grouping [3].
• Multiple Sequence Alignment: Perform alignment using MAFFT 7.0 or Clustal Omega [3] [19].
• Phylogenetic Reconstruction: Construct maximum likelihood trees using FastTreeMP or MEGA with 1000 bootstrap replicates [3] [19].
• Duplication Analysis: Identify tandem duplication events by analyzing genomic clustering patterns [3].

Experimental Workflow for Comprehensive NBS Gene Analysis

Expression Profiling and Functional Validation

Protocol 3: Transcriptomic Analysis of NBS Genes

• Data Retrieval: Obtain RNA-seq data from specialized databases (IPF database, CottonFGD, Cottongen) and NCBI BioProjects [3].
• Data Categorization: Classify expression data into tissue-specific, abiotic stress-specific, and biotic stress-specific categories [3].
• Expression Quantification: Extract FPKM values and process through transcriptomic pipelines [3].
• Visualization: Generate heat maps to visualize differential expression patterns across conditions [3].

Protocol 4: Functional Validation through Genetic Approaches

• Genetic Variation Analysis: Identify sequence variants between susceptible and tolerant accessions through whole-genome comparisons [3].
• Protein Interaction Studies: Conduct protein-ligand and protein-protein interaction assays to validate interactions with pathogen effectors [3].
• Functional Characterization: Implement virus-induced gene silencing (VIGS) to assess gene function in resistant varieties [3].
• Phenotypic Assessment: Evaluate disease symptoms and pathogen titers in silenced plants [3].

Signaling Pathways and Molecular Mechanisms

NBS-LRR Activation and Signal Transduction

NBS-LRR proteins function as molecular switches in plant immunity, transitioning between ADP-bound inactive states and ATP-bound active states [14] [22]. The NBS domain facilitates nucleotide binding and hydrolysis, with the LRR domain implicated in both effector recognition and intramolecular interactions [22].

Research on the potato Rx protein (a CNL) revealed that intramolecular interactions between domains maintain the protein in an autoinhibited state in the absence of pathogen elicitors [22]. Pathogen recognition induces conformational changes through sequential disruption of these interactions, leading to activation [22]. Specifically, the Rx protein exhibits interactions between its CC and NBS-LRR domains that are disrupted in the presence of the potato virus X coat protein elicitor [22].

NBS-LRR Protein Activation Pathway

Regulatory Networks

The expression of NBS-LRR genes is tightly regulated due to the potential fitness costs associated with their inappropriate activation [21] [19]. MicroRNA-mediated regulation represents a crucial layer of control, with diverse miRNA families (e.g., miR482/2118) targeting conserved NBS-LRR motifs [21]. These miRNAs typically target highly duplicated NBS-LRRs, while heterogeneous NBS-LRR families are less frequently targeted [21].

Analyses of NLR genes in asparagus species revealed their promoters contain numerous cis-elements responsive to defense signals and phytohormones, indicating complex transcriptional regulation [19]. Domesticated species like A. officinalis show both contraction of NLR gene repertoire and reduced induction of retained NLR genes compared to wild relatives, suggesting artificial selection has impacted regulatory networks [19].

The Scientist's Toolkit: Essential Research Reagents

Table 3: Key Research Reagents for NBS Gene Analysis

Reagent/Resource	Specific Example	Application	Function	Reference
Genome Databases	NCBI, Phytozome, Plaza	Gene identification	Source of genome assemblies and annotations	[3]
Domain Databases	Pfam, InterProScan	Domain architecture analysis	Identification of NBS and associated domains	[3] [19]
HMM Models	Pfam-A_hmm model	Domain screening	Identification of NB-ARC domains	[3]
Orthology Tools	OrthoFinder v2.5.1	Evolutionary analysis	Orthogroup construction and classification	[3]
Phylogenetic Software	FastTreeMP, MEGA	Evolutionary analysis	Phylogenetic tree construction	[3] [19]
Expression Databases	IPF database, CottonFGD	Expression profiling	Source of RNA-seq data	[3]
Promoter Analysis Tools	PlantCARE	Regulatory element identification	Prediction of cis-acting regulatory elements	[19]
Motif Analysis Tools	MEME Suite	Conserved motif prediction	Identification of conserved NBS motifs	[19]
VIGS Systems	Virus-induced gene silencing	Functional validation	Transient gene silencing in plants	[3]
12alpha-Fumitremorgin C	12alpha-Fumitremorgin C, MF:C22H25N3O3, MW:379.5 g/mol	Chemical Reagent	Bench Chemicals
Boc-DL-Arg(Pmc)(Pmc)-OH	Boc-DL-Arg(Pmc)(Pmc)-OH, MF:C25H40N4O7S, MW:540.7 g/mol	Chemical Reagent	Bench Chemicals

The comparative analysis of NBS domain genes across plant species reveals both conserved evolutionary patterns and lineage-specific innovations. Bryophytes possess unexpected diversity with novel NBS classes like PNL and HNL, while angiosperms exhibit dynamic expansions particularly following the Cretaceous-Paleogene boundary. The structural and functional conservation of NBS domains across 500 million years of plant evolution underscores their fundamental role in plant immunity.

Future research directions should include comprehensive functional characterization of bryophyte-specific NBS classes, exploration of regulatory networks controlling NBS gene expression, and utilization of comparative genomic insights for crop improvement. The experimental methodologies and resources outlined in this guide provide a foundation for advancing our understanding of plant immunity mechanisms across the evolutionary spectrum.

Gene duplication is a fundamental engine of evolutionary innovation, providing the raw genetic material for the evolution of new functions and adaptive traits. Among the various mechanisms of gene duplication, whole-genome duplication (WGD) and tandem duplication (TD) represent two fundamentally distinct processes with profound implications for genome evolution and gene content [23]. WGD involves the duplication of an entire genome, creating massive genetic redundancy across all loci, while TD generates localized clusters of duplicated genes through the repeated copying of individual genes or genomic segments [24]. Understanding the relative contributions and evolutionary consequences of these duplication mechanisms is particularly crucial for interpreting the expansion and diversification of key gene families, such as the nucleotide-binding site (NBS) domain genes that comprise the majority of plant disease resistance (R) genes [3] [8]. This guide provides a comparative analysis of WGD and TD, synthesizing current genomic evidence to elucidate their distinct roles in shaping plant genome architecture, functional diversity, and adaptive potential.

Comparative Analysis of Duplication Mechanisms

Fundamental Characteristics and Genomic Signatures

Table 1: Fundamental Characteristics of Whole-Genome and Tandem Duplication

Feature	Whole-Genome Duplication (WGD)	Tandem Duplication (TD)
Genomic Scale	Entire genome duplication [23]	Single genes or small genomic segments [23]
Frequency of Occurrence	Episodic, rare events (every ~10-100 million years) [23]	Continuous, frequent events [23]
Typical Gene Copy Number	All genes doubled in a single event [24]	2 or more copies in close proximity [25]
Genomic Distribution	Genome-wide, creating systemic blocks [23]	Localized clusters on specific chromosomes [23]
Evolutionary Half-Life	Long-term retention of some duplicates [26]	Short half-life, rapid turnover [23]
Inheritance Pattern	All genes duplicated simultaneously	Gene-by-gene basis

Evolutionary Dynamics and Functional Consequences

The differential mechanisms of WGD and TD impose distinct selective pressures and evolutionary trajectories on their duplicated products, leading to significant functional biases in gene retention and diversification.

Table 2: Evolutionary Outcomes of Whole-Genome and Tandem Duplication

Evolutionary Parameter	Whole-Genome Duplication (WGD)	Tandem Duplication (TD)
Primary Functional Bias	Dosage-sensitive genes, transcription factors, core cellular processes [26] [25]	Environmental response genes, biotic/abiotic stress resistance [24] [25]
Typical Expression Divergence	Gradual subfunctionalization or conservation of broad expression [26]	Rapid neofunctionalization or asymmetric expression [26]
Selection Pressure	Weaker purifying selection, especially initially [26]	Stronger selective pressure [23]
Retention of Redundant Copies	High for dosage-sensitive genes [26]	Low, rapid functional divergence or loss [23]
Role in Adaptation	Major genomic revolutions, morphological innovation [23]	Continuous adaptation to rapidly changing environments [24] [23]
Impact on NBS Gene Evolution	Large-scale expansion followed by fractionation [8]	Species-specific, lineage-specific expansion of R-genes [3] [8]

The relationship between duplication mechanism and gene function is particularly striking. WGD-derived genes are preferentially retained for dose-sensitive genes involved in essential cellular processes like DNA-binding, transcription factor activity, and core metabolism [26] [25]. This retention bias is explained by the gene balance hypothesis, which predicts that components of multiprotein complexes require stoichiometric balance [26]. In contrast, TD-derived genes are overwhelmingly enriched for functions in environmental interactions, particularly defense responses against pathogens and abiotic stresses [24]. This functional specialization makes TD a critical mechanism for the rapid expansion of disease resistance gene families, including NBS-encoding genes [3].

Figure 1: Evolutionary trajectories of gene duplicates following whole-genome versus tandem duplication. WGD and TD produce duplicates with distinct functional biases and evolutionary fates, shaping genome evolution through complementary mechanisms.

Experimental Approaches for Studying Duplication Events

Genomic and Bioinformatics Workflows

Research in this field relies on integrated genomic, transcriptomic, and bioinformatic approaches to identify duplication events and characterize their functional consequences.

Table 3: Key Experimental and Bioinformatics Methodologies

Methodology	Primary Application	Key Insights Generated
Synteny Analysis	Identifying WGD-derived genomic blocks [23]	Reveals ancient polyploidization events and systemic relationships
Ks Distribution Analysis	Dating duplication events [23]	Identifies peaks of duplication events in evolutionary history
Hidden Markov Model (HMM) Profiling	Identifying NBS domain genes [3] [8]	Enables genome-wide identification of resistance gene families
OrthoFinder/OrthoMCL	Classifying orthologous groups [3]	Distinguishes lineage-specific expansion from shared gene families
RNA-seq Expression Profiling	Characterizing expression divergence [26] [3]	Reveals subfunctionalization and neofunctionalization patterns
Virus-Induced Gene Silencing (VIGS)	Functional validation of candidate genes [3]	Tests role of specific duplicates in disease resistance

Figure 2: Experimental workflow for studying duplication events and their functional consequences. Integrated genomic and functional approaches enable comprehensive characterization of WGD and TD events and their roles in evolution.

The Scientist's Toolkit: Essential Research Reagents

Table 4: Essential Research Reagents and Resources for Studying Gene Duplication

Resource Category	Specific Examples	Function and Application
Genomic Databases	Phytozome, BRAD, Bolbase, PLAZA [3] [8]	Provide annotated genome sequences and comparative genomics tools
Domain Databases	Pfam (PF00931 for NBS domain) [3] [8]	Hidden Markov Models for identifying protein domains
Bioinformatics Tools	DupGen_finder, OrthoFinder, DIAMOND, MCLE [3] [23]	Identify and classify duplication modes and orthologous groups
Expression Databases	IPF Database, CottonFGD, NCBI BioProjects [3]	Provide RNA-seq data for expression divergence analysis
Functional Validation Tools	Virus-Induced Gene Silencing (VIGS) vectors [3]	Enable functional characterization of duplicated genes
APA amoxicillin amide	APA Amoxicillin Amide	Get high-quality APA Amoxicillin Amide (CAS 1789703-32-7), a key impurity for amoxicillin research. For Research Use Only. Not for human or veterinary use.
Oripavine-d3	Oripavine-d3|Stable Isotope|For Research	Oripavine-d3 is a deuterated internal standard for precise quantification of oripavine in research. This product is for Research Use Only and not for human or veterinary diagnosis or therapeutic use.

Case Study: NBS Domain Gene Evolution in Plants

The evolution of NBS-encoding disease resistance genes provides an excellent model for understanding the complementary roles of WGD and TD. These genes are crucial for plant immunity and exhibit remarkable diversity across plant species.

Duplication Patterns in Brassica Species

Comparative analysis of NBS-encoding genes in Brassica oleracea, Brassica rapa, and Arabidopsis thaliana reveals a complex evolutionary history shaped by both WGD and TD [8]. The Brassica lineage experienced a whole-genome triplication (WGT) event after its divergence from Arabidopsis ~16 million years ago [8]. Following this WGT event, NBS-encoding homologous gene pairs on triplicated regions were rapidly deleted or lost. However, subsequent species-specific gene amplification occurred through TD, leading to the expansion of NBS gene families in each lineage [8]. This pattern demonstrates how large-scale WGD events can provide genetic raw material that is subsequently refined and specialized through small-scale TD events.

Expression Divergence and Functional Specialization

Spatial transcriptomics technologies have revealed that the mechanism of duplication profoundly influences expression divergence between paralogs [26]. Duplication mechanisms that preserve cis-regulatory landscapes, such as WGD and TD, typically yield paralogs with more conserved expression profiles [26]. However, over time, TD-derived genes often diverge asymmetrically, with one copy maintaining broad expression while the other specializes in specific cell types or conditions [26]. This expression specialization is particularly relevant for NBS-encoding genes, which may evolve new specificities against rapidly evolving pathogens through TD-mediated expansion [3].

Recent research on NBS-encoding genes in cotton demonstrated how tandemly duplicated orthogroups (OG2, OG6, and OG15) show putative upregulation in different tissues under various biotic and abiotic stresses [3]. Functional validation through virus-induced gene silencing (VIGS) of a candidate gene (GaNBS from OG2) confirmed its role in virus resistance, illustrating the adaptive significance of TD-derived NBS genes [3].

Evolutionary Implications and Future Directions

The complementary actions of WGD and TD have shaped plant genome evolution through distinct but interconnected mechanisms. WGD events provide evolutionary revolutions—cataclysmic genomic changes that create massive genetic redundancy and enable major functional innovations over long evolutionary timescales [23]. In contrast, TD provides continuous evolutionary tinkering—a steady supply of genetic variants that enable fine-tuned adaptations to rapidly changing environmental conditions, especially in stress response pathways [24] [23].

This duality is particularly evident in the evolution of plant immune systems. WGD events have created large reservoirs of genetic material that can be co-opted for disease resistance functions, while TD enables the rapid, lineage-specific expansion of resistance gene families in response to emerging pathogen threats [3] [8]. The functional specialization of TD-derived genes for environmental interactions makes this mechanism particularly important for adaptive evolution in rapidly changing environments [24].

Future research directions include leveraging spatial transcriptomics to understand expression divergence at cellular resolution [26], exploring the role of epigenetic modifications in duplicate gene regulation, and investigating how duplication mechanisms influence protein interaction networks and metabolic pathways. Understanding these evolutionary dynamics has practical implications for crop improvement, suggesting that manipulating both WGD (through synthetic polyploidy) and TD (through gene editing) may provide strategies for enhancing disease resistance and environmental resilience in agricultural systems.

Plant disease resistance (R) genes are a key component of the innate immune system that protects plants from a diverse range of pathogens. The nucleotide-binding site (NBS) gene family represents one of the largest classes of R genes, encoding proteins that play critical roles in effector-triggered immunity (ETI). These proteins typically contain an NB-ARC (nucleotide-binding adaptor shared by APAF-1, R proteins, and CED-4) domain and are often accompanied by C-terminal leucine-rich repeats (LRRs) and variable N-terminal domains such as TIR (Toll/Interleukin-1 receptor), CC (coiled-coil), or RPW8 (Resistance to Powdery Mildew 8). Based on these domain architectures, NBS-encoding genes are classified into several types including TNL (TIR-NBS-LRR), CNL (CC-NBS-LRR), RNL (RPW8-NBS-LRR), and various truncated forms lacking complete domain suites [27] [3].

The distribution and evolution of NBS-encoding genes vary considerably across plant species. In flowering plants, substantial gene expansion has occurred, resulting in extensive NBS repertoires. For instance, the ANNA database documents over 90,000 NLR genes from 304 angiosperm genomes, including 18,707 TNL genes, 70,737 CNL genes, and 1,847 RNL genes. This stands in stark contrast to bryophytes like Physcomitrella patens, which possess only around 25 NLRs, suggesting that significant gene expansion occurred primarily in flowering plants [3].

This case study examines the genomic expansion of NBS domain genes in cotton (Gossypium species) and peanut (Arachis species), two economically important crops with distinct evolutionary histories. Through comparative analysis, we explore how differential expansion patterns and evolutionary trajectories of NBS genes have influenced disease resistance profiles in these crop species.

Genomic Distribution of NBS Genes in Cotton and Peanut

NBS Gene Profiles in Cotton Species

Comparative genomic analyses reveal significant variation in NBS-encoding gene content across cotton species. Studies have identified 246, 365, 588, and 682 NBS-encoding genes in G. arboreum (A-genome), G. raimondii (D-genome), G. hirsutum (allotetraploid), and G. barbadense (allotetraploid), respectively. The distribution of these genes among chromosomes is nonrandom and uneven, with a tendency to form clusters. Notably, the two allotetraploid cotton species possess approximately twice the number of NBS genes compared to their diploid progenitors, suggesting preservation and potential expansion following hybridization [27].

Domain architecture analysis shows substantial differences between cotton species. G. arboreum and G. hirsutum possess a greater proportion of CN (CC-NBS), CNL, and N (NBS-only) genes, and a lower proportion of NL (NBS-LRR), TN (TIR-NBS), and TNL genes compared to G. raimondii and G. barbadense. The most dramatic difference is observed in TNL genes, with G. raimondii and G. barbadense containing approximately seven times the percentage of TNL genes found in G. arboreum and G. hirsutum. This asymmetric distribution has functional implications, as TNL genes are associated with resistance to Verticillium wilt [27].

Table 1: NBS-Encoding Gene Distribution in Cotton Species

Species	Genome Type	Total NBS Genes	CN (%)	CNL (%)	N (%)	NL (%)	TNL (%)	Other (%)
G. arboreum	Diploid (A)	246	17.89	32.52	23.98	21.54	2.03	2.04
G. raimondii	Diploid (D)	365	10.68	29.32	16.99	24.38	13.70	4.93
G. hirsutum	Allotetraploid (AD)	588	15.14	28.06	28.57	26.19	0.85	1.19
G. barbadense	Allotetraploid (AD)	682	13.49	20.97	25.07	30.79	6.45	3.23

NBS Gene Profiles in Peanut Species

Peanut exhibits a different pattern of NBS gene expansion. In cultivated peanut (A. hypogaea cv. Tifrunner), 713 full-length NBS-LRR genes have been identified, with 229 containing TIR domains, 118 containing CC domains, and surprisingly, 26 sequences containing both TIR and CC domains—a feature not observed in the diploid progenitors. This suggests that genetic exchange or gene rearrangement likely resulted in domain fusion after tetraploidization [28].

Wild peanut species show distinct NBS gene profiles. Studies have identified 393 and 437 NBS-LRR genes in A. duranensis (A-genome) and A. ipaensis (B-genome), respectively. Among these, 278 and 303 were full-length sequences. Comparative analysis revealed that A. ipaensis has more gene clusters than A. duranensis, possibly due to more frequent tandem duplication events. The LRR domains in these genes mainly underwent purifying selection, though most LRR8 domains experienced positive selection, suggesting adaptive evolution [29].

Table 2: NBS-Encoding Gene Distribution in Peanut Species

Species	Genome Type	Total NBS Genes	Full-Length Genes	TNL (%)	CNL (%)	TNL+CNL (%)	NBS-WRKY Fusion
A. duranensis	Diploid (A)	393	278	32.1	67.9	0	Not reported
A. ipaensis	Diploid (B)	437	303	31.4	68.6	0	Not reported
A. hypogaea	Allotetraploid (AB)	713	713	32.1	67.3	0.36	3 genes

Evolutionary Dynamics and Selection Pressures

Evolutionary Patterns in Cotton

Phylogenetic analysis in cotton reveals that TIR-NBS genes of G. barbadense are closely related to those of G. raimondii, while G. hirsutum shows greater similarity to G. arboreum. Synteny analysis supports this pattern, indicating that G. hirsutum inherited more NBS-encoding genes from G. arboreum, while G. barbadense inherited more from G. raimondii. This asymmetric evolution of NBS-encoding genes may explain differential disease resistance between these species [27].

Notably, G. raimondii and G. barbadense demonstrate higher resistance to Verticillium wilt, while G. arboreum and G. hirsutum are more susceptible. This correlation suggests that TNL genes, which are more abundant in Verticillium-resistant species, may play significant roles in resistance to this pathogen. The differences in NBS gene repertoire between tetraploid cottons and their diploid progenitors indicate that allopolyploidization was followed by either preferential gene retention from one progenitor or differential gene loss [27].

Evolutionary Patterns in Peanut

In peanut, evolutionary analysis reveals that NBS-LRR proteins and LRR domains have undergone relaxed selection in cultivated peanut compared to wild diploids. Particularly noteworthy is the preferential loss of LRR domains in cultivated peanut, which may partially explain its generally lower disease resistance compared to wild relatives. Despite this trend, quantitative trait locus (QTL) analysis has identified 113 NBS-LRRs associated with response to late leaf spot, tomato spotted wilt virus, and bacterial wilt in cultivated peanut [28].

These resistance-associated NBS-LRRs in cultivated peanut were classified as 75 young and 38 old genes, indicating that young NBS-LRRs produced after tetraploidization play significant roles in disease resistance. This finding highlights the importance of recent gene evolution in adapting to pathogen pressures. The pangenome analysis of peanut further revealed substantial structural variations affecting NBS genes, with 1,335 domestication-related structural variations and 190 structural variations associated with seed size or weight identified [30] [28].

Methodologies for NBS Gene Identification and Analysis

Standard Bioinformatics Pipeline for NBS Gene Identification

The identification and classification of NBS-encoding genes typically follows a standardized bioinformatics workflow. First, genome sequences or protein databases are searched using Hidden Markov Model (HMM) profiles corresponding to the NB-ARC domain (PF00931) from the Pfam database. The HMMER software package is commonly employed with default e-value cutoffs (often 1.1e-50) to ensure stringent selection [3] [29].

Following initial identification, additional domains (TIR, CC, RPW8, LRR) are detected using complementary approaches:

• Pfam and SMART databases for TIR and LRR domains
• MARCOIL or Paircoil2 programs for CC domains (P-score cutoff typically 0.03)
• Custom parsing scripts to classify genes into architectural types based on domain combinations [27] [31] [29]

Evolutionary and Expression Analysis Methods

For evolutionary analysis, multiple sequence alignment of full-length protein sequences is performed using tools such as MAFFT or ClustalW with default parameters. Phylogenetic trees are constructed using maximum likelihood (ML) or neighbor-joining (NJ) methods implemented in MEGA or similar software, with bootstrap validation (typically 1000 replicates) [3] [29].

Selection pressure is assessed by calculating nonsynonymous (Ka) and synonymous (Ks) substitution rates using PAML or similar packages. Ka/Ks ratios >1, =1, and <1 indicate positive, neutral, and purifying selection, respectively [29].

Gene expression analysis under pathogen challenge typically involves:

• Pathogen inoculation under controlled conditions
• RNA extraction at multiple time points
• qRT-PCR with gene-specific primers
• Reference gene normalization (e.g., actin)
• Statistical analysis of expression differences [29]

Disease Resistance Associations

Cotton Resistance Profiles and NBS Gene Correlations

Cotton species show distinct resistance patterns that correlate with their NBS gene profiles. Verticillium wilt, caused by the soilborne fungal pathogen Verticillium dahliae, presents a particularly clear example of this relationship. G. raimondii is nearly immune to this pathogen, and G. barbadense is typically resistant or highly resistant, whereas G. arboreum and G. hirsutum are often susceptible. This resistance pattern strongly correlates with the abundance of TNL genes, which are significantly more prevalent in resistant species [27].

For Fusarium wilt, caused by Fusarium oxysporum, the resistance pattern differs. G. barbadense is often more susceptible to F. oxysporum compared to G. arboreum and G. hirsutum, indicating that different NBS gene types may confer resistance to different pathogens [27].

Analysis of the correlation between disease resistance QTL and NBS-encoding genes in G. raimondii suggests that more than half of disease resistance QTL are associated with NBS-encoding genes. This agrees with previous studies establishing that more than half of plant resistance genes are NBS-encoding genes [31].

Peanut Resistance Profiles and NBS Gene Correlations

In peanut, resistance to various pathogens has been associated with NBS-LRR genes. In A. duranensis, A. ipaensis, and A. hypogaea cv. Tifrunner, NBS-LRRs have been identified within QTL regions responsive to late leaf spot, tomato spotted wilt virus, and bacterial wilt. Specifically, 2, 39, and 113 NBS-LRRs were associated with these diseases in the respective species [28].

Expression profiling following Aspergillus flavus infection revealed differential expression patterns between wild and cultivated peanuts. In A. duranensis, upregulated expression of NBS-LRR genes was continuous after infection, while these genes responded temporally in cultivated peanut (A. hypogaea). This temporal expression pattern in cultivated peanut may contribute to its greater susceptibility to A. flavus infection and subsequent aflatoxin contamination [29].

Recent functional validation using virus-induced gene silencing (VIGS) demonstrated that silencing of a specific NBS gene (GaNBS, OG2) in resistant cotton reduced its resistance, confirming the functional role of NBS genes in disease resistance [3].

Research Reagent Solutions Toolkit

Table 3: Essential Research Reagents and Tools for NBS Gene Analysis

Category	Specific Tool/Reagent	Function/Application	Example Use
Bioinformatics Tools	HMMER v3.1b2	Domain-based gene identification	Identifying NB-ARC domains in genome assemblies [27]
	Pfam Database	Protein family annotation	Verifying NBS, TIR, LRR domains [3]
	SMART	Protein domain analysis	Detecting functional domains in NBS proteins [31]
	MARCOIL/Paircoil2	Coiled-coil domain prediction	Identifying CC domains in NBS proteins [31] [29]
	OrthoFinder	Orthogroup inference	Clustering NBS genes across species [3]
Evolutionary Analysis	MAFFT v7.0	Multiple sequence alignment	Aligning NBS protein sequences [3] [29]
	MEGA v6.0/5.05	Phylogenetic analysis	Constructing evolutionary trees [31] [29]
	PAML v4.0	Selection pressure analysis	Calculating Ka/Ks ratios [29]
Experimental Validation	Virus-Induced Gene Silencing (VIGS)	Functional characterization	Validating NBS gene function in resistance [3]
	qRT-PCR with SYBR Green	Expression profiling	Measuring NBS gene expression under pathogen challenge [29]
Ethyl propargyl sulfone	Ethyl Propargyl Sulfone|Research Use Only	Ethyl Propargyl Sulfone is a versatile building block for synthesizing bioactive cyclic sulfones. This product is for research purposes only and not for human use.	Bench Chemicals

This case study reveals both convergent and divergent patterns in NBS gene expansion between cotton and peanut. Both crops show significant expansion of NBS genes in their allotetraploid forms compared to diploid progenitors, yet the specific evolutionary trajectories and functional outcomes differ substantially.

In cotton, asymmetric evolution following allopolyploidization resulted in species-specific NBS gene profiles that correlate with differential disease resistance. The inheritance patterns from diploid progenitors to allotetraploid descendants significantly influenced resistance capabilities, particularly against Verticillium wilt. The abundance of TNL genes emerged as a key factor in Verticillium resistance [27].

In peanut, the evolutionary story is characterized by relaxed selection on NBS-LRR proteins and preferential loss of LRR domains in cultivated varieties, potentially explaining their generally lower disease resistance compared to wild relatives. Despite this trend, the production of young NBS-LRR genes after tetraploidization appears crucial for maintaining disease resistance capabilities. The discovery of genes with both TIR and CC domains in cultivated peanut, but not in diploid progenitors, highlights the ongoing evolution and innovation in the NBS gene family following polyploidization [28].

These comparative genomic analyses provide valuable insights for crop improvement strategies. Understanding the specific NBS gene architectures associated with disease resistance in these crops enables more targeted breeding approaches and genetic engineering strategies to enhance disease resistance while maintaining favorable agronomic traits.

Advanced Computational and Experimental Methods for NBS Gene Discovery

This guide provides a comparative analysis of three foundational bioinformatics tools—HMMER, Pfam, and OrthoFinder—within the context of comparative genomics research on Nucleotide-Binding Site (NBS) domain genes in plants. The evaluation, grounded in experimental data from recent studies, demonstrates that an integrated pipeline leveraging these tools enables high-accuracy domain identification, orthogroup inference, and evolutionary analysis, providing critical insights into plant disease resistance gene families.

The table below summarizes the core functionality and typical usage of each tool in a comparative genomics workflow.

Tool	Primary Function	Role in Comparative Genomics	Methodology
HMMER	Profile Hidden Markov Model (HMM) search for sensitive sequence homology detection [32]	Identifies protein domains (e.g., NB-ARC) in query sequences against domain databases like Pfam [3].	Statistical probability models for detecting remote homologs.
Pfam	Curated database of protein families and domains [33]	Provides the HMM profiles (e.g., PF00931 for NB-ARC) used by HMMER to annotate domains in gene sets [3].	Large collection of multiple sequence alignments and HMMs.
OrthoFinder	Phylogenetic orthology inference from whole proteomes [34]	Clusters genes into orthogroups, infers gene trees and species trees, and identifies gene duplication events [3].	Graph-based clustering (orthogroups) and phylogenetic tree analysis.

Performance Comparison and Benchmarking Data

Orthology Inference Accuracy

OrthoFinder has been extensively benchmarked against other methods. The table below summarizes its performance on the Quest for Orthologs benchmark, a community-standardized evaluation [34].

Method	Ortholog Inference Accuracy (SwissTree Test)	Ortholog Inference Accuracy (TreeFam-A Test)	Key Strengths
OrthoFinder (Default)	3-24% higher than other methods [34]	2-30% higher than other methods [34]	Most accurate ortholog inference; provides comprehensive phylogenetic outputs [34].
Other Methods (e.g., InParanoid, OrthoMCL, OMA)	Lower accuracy range [34]	Lower accuracy range [34]	Varying strengths, but none consistently second best [34].

A key reason for OrthoFinder's high accuracy is its phylogenetic approach, which uses gene trees to distinguish orthologs from paralogs, overcoming limitations of score-based heuristic methods that can be confounded by variable sequence evolution rates [34].

Application Performance in NBS Gene Research

A 2024 study on NBS genes in plants utilized a pipeline integrating these tools. The following table summarizes the scale and performance of this integrated approach [3].

Performance Metric	Result	Context and Implication
Genomes Analyzed	34 plant species [3]	Broad taxonomic coverage from mosses to monocots and dicots.
NBS Genes Identified	12,820 genes [3]	Demonstrates HMMER/Pfam's scalability for large-scale genome annotation.
Domain Architecture Classes	168 classes identified [3]	Pfam-based domain annotation reveals extensive functional diversity.
Orthogroups (OGs) Clustered	603 OGs with OrthoFinder [3]	Effective delineation of evolutionary lineages; identified core and species-specific OGs.

Experimental Protocols for NBS Domain Gene Analysis

The following workflow, based on a published study [3], details a standard protocol for the comparative analysis of a gene family across multiple species.

Step 1: Identification of NBS Domain-Containing Genes

• Objective: To identify all genes containing the NBS (NB-ARC) domain in a set of proteomes.
• Method:
  • Tool: HMMER3 (specifically PfamScan.pl).
  • Database: Pfam-A.hmm model for the NB-ARC domain (PF00931).
  • Parameters: A strict E-value cutoff of 1.1e-50 is used to ensure high-confidence matches [3].
  • Output: A list of all genes containing the NBS domain for each species.

Step 2: Domain Architecture Classification

• Objective: To classify the identified NBS genes based on their full domain composition.
• Method:
  • All additional domains (e.g., TIR, LRR, CC) associated with the NBS genes are identified using the same HMMER/Pfam workflow.
  • Genes are classified into architectural classes based on the combination and order of their domains [3].
  • Output: A comprehensive classification of classical (e.g., TIR-NBS-LRR) and species-specific structural patterns.

Step 3: Evolutionary Analysis and Orthogroup Inference

• Objective: To cluster NBS genes into orthogroups (groups of genes descended from a single gene in the last common ancestor of the species being compared) to understand evolutionary relationships.
• Method:
  • Tool: OrthoFinder v2.5.1.
  • Sequence Search: The DIAMOND tool is used for fast all-vs-all sequence similarity searches [3].
  • Clustering: The MCL algorithm clusters genes into orthogroups based on the similarity scores [3].
  • Output: Orthogroups (OGs), which can be categorized as "core" (common across many species) or "unique" (species-specific).

Step 4: Phylogenetic and Duplication Analysis

• Objective: To infer evolutionary relationships and identify gene duplication events.
• Method:
  • Tool: OrthoFinder (internal workflow).
  • Alignment & Tree Building: OrthoFinder uses MAFFT for multiple sequence alignment and FastTreeMP for maximum likelihood gene tree construction within orthogroups [3].
  • Duplication Inference: OrthoFinder analyzes gene trees and the species tree to identify gene duplication events [3].
  • Output: Rooted gene trees, a rooted species tree, and a list of gene duplication events.

The following diagram visualizes this integrated experimental workflow.

Diagram 1: Integrated bioinformatics pipeline for comparative analysis of NBS domain genes.

This table lists key databases, tools, and resources essential for conducting research in this field.

Resource Name	Type	Function in the Pipeline
Pfam Database [35] [33]	Protein Family Database	Provides the curated HMM profiles for identifying protein domains like the NB-ARC domain [3].
DIAMOND [34] [3]	Sequence Similarity Search Tool	A faster alternative to BLAST for all-vs-all sequence searches, used by OrthoFinder for initial similarity comparisons [3].
MAFFT [3]	Multiple Sequence Alignment Tool	Used for creating accurate alignments of protein sequences within orthogroups for phylogenetic analysis [3].
FastTreeMP [3]	Phylogenetic Tree Inference Tool	Used for inferring approximate maximum-likelihood gene trees from multiple sequence alignments [3].
EggNOG [36] [34]	Orthology Database	A public database of orthologous groups and functional annotation, useful for comparison and validation [37].

The integrated use of HMMER/Pfam for precise domain annotation and OrthoFinder for phylogenetic orthology inference creates a powerful and accurate pipeline for comparative genomic studies. Benchmarking data confirms that OrthoFinder outperforms other methods in ortholog detection accuracy, while real-world application in plant NBS gene research demonstrates the pipeline's robustness and scalability. This combination of tools enables researchers to reliably uncover evolutionary patterns and functional diversification in gene families critical for traits like disease resistance.

Machine Learning and Deep Learning Classifiers for R-protein Prediction

Plant resistance genes (R-genes), particularly those encoding nucleotide-binding site leucine-rich repeat (NBS-LRR or NLR) proteins, constitute a primary line of defense in the plant immune system, enabling recognition of pathogen effectors and initiation of effector-triggered immunity (ETI) [38] [39]. The identification and classification of these genes are critical for understanding plant defense mechanisms and for breeding disease-resistant crops. Traditional methods for R-gene identification, which rely on sequence similarity and domain search tools like BLAST, HMMER, and InterProScan, often struggle with the immense diversity and rapid evolution of these genes, frequently missing novel or highly divergent sequences [38]. The advent of machine learning (ML) and deep learning (DL) has begun to transform this field, offering powerful, alignment-free methods for the accurate prediction and classification of R-genes from sequence data alone. This guide provides a comparative analysis of contemporary computational classifiers for R-protein prediction, situating them within the broader research context of comparative NBS domain gene analysis across plant species [3]. We objectively evaluate the performance, underlying methodologies, and practical applications of these tools to assist researchers in selecting the most appropriate solutions for their work.

Traditional Methods and the Need for Advanced Classifiers

Conventional R-gene Identification Pipelines

Before the rise of ML/DL approaches, the standard pipeline for identifying NBS-LRR genes involved a multi-step process. Researchers typically began with a genome-wide search using tools like HMMER3 with Hidden Markov Models (HMMs) of the NB-ARC domain (PF00931) or performing BLASTP searches with known NBS sequences as queries [3] [39] [40]. Candidate genes were then subjected to domain analysis using PfamScan, NCBI-CDD, or SMART to confirm the presence of characteristic domains such as TIR, CC, LRR, and RPW8 [3] [40]. Finally, gene classification into subfamilies (e.g., TNL, CNL, RNL) was performed based on the combination of identified domains [39] [40].

Limitations of Traditional Approaches

While effective, these homology-based methods possess significant limitations. They often produce fragmented annotations due to the complex genomic structure of R-gene clusters and their tendency to be misidentified as repetitive elements [38]. Their performance drops considerably when sequence similarity to known R-genes is low, making them poorly suited for discovering novel R-gene classes in newly sequenced or non-model plant genomes [38]. The manual curation required to validate results is time-consuming and not scalable for large genomic studies.

Comparative Performance of ML/DL Classifiers

The following table summarizes the performance metrics of leading ML/DL-based R-gene prediction tools as reported in their respective studies.

Table 1: Performance Comparison of R-gene Prediction Tools

Tool Name	Underlying Algorithm	Primary Function	Reported Accuracy	Key Advantages
PRGminer [38]	Deep Learning (Dipeptide Composition)	R-gene identification & classification into 8 classes	98.75% (k-fold), 95.72% (independent test)	High accuracy with MCC of 0.98; webserver available
DPFunc [41]	Deep Learning (GCN with Domain-guided Attention)	Protein function prediction, incl. defense response	Significant improvement over SOTA (Fmax: 16-27% increase)	Integrates domain info for interpretability; detects key functional residues
PCPIP [42]	Support Vector Machine (SVM)	Classification of native vs. non-native PPI interfaces	High performance on benchmarking datasets	Effective for identifying biologically relevant protein complexes

Detailed Methodologies of Featured Classifiers

PRGminer: A Deep Learning Framework for R-gene Prediction

PRGminer is a dedicated DL tool designed specifically for the high-throughput prediction of plant R-genes. Its implementation occurs in two distinct phases [38].

Phase I: R-gene vs. Non-R-gene Classification

• Input Representation: The tool converts protein sequences into a numerical representation using dipeptide composition, which was found to yield superior performance compared to other sequence encoding methods.
• Model Architecture: A deep learning model processes this input to classify sequences as either R-genes or non-R-genes. The model achieves an accuracy of 98.75% in a k-fold training/testing procedure and 95.72% on an independent test set, with a high Matthews Correlation Coefficient (MCC) of 0.98 and 0.91, respectively [38].

Phase II: R-gene Subclassification

• Process: Sequences predicted as R-genes in Phase I are subsequently classified into one of eight categories: CNL, TNL, TIR, RLK, RLP, LECRK, LYK, and KIN. This multi-classification achieved an overall accuracy of 97.55% in k-fold testing and 97.21% on an independent set [38].
• Availability: PRGminer is accessible as both a user-friendly webserver and a standalone tool, facilitating adoption by researchers with varying computational expertise.

DPFunc: Leveraging Domain Guidance for Functional Prediction

While not exclusively an R-gene predictor, DPFunc is a state-of-the-art DL model for general protein function prediction that can be powerfully applied to identify proteins involved in defense responses. Its methodology is notable for its integration of structural and domain information [41].

Workflow:

• Residue-level Feature Learning: The protein sequence is passed through a pre-trained protein language model (ESM-1b) to generate initial residue-level features. A protein structure-based contact map is constructed (from experimental or AlphaFold-predicted structures), and Graph Convolutional Network (GCN) layers propagate and update features between residues [41].
• Domain-guided Attention: Domain information is detected from the sequence using InterProScan. An attention mechanism, inspired by transformer architectures, then uses these domain embeddings to guide the model toward functionally crucial residues in the protein structure, creating an interpretable link between domain, structure, and function [41].
• Function Prediction: The weighted residue features are aggregated into a protein-level representation, which is used to predict Gene Ontology (GO) terms, including those related to immune responses.

Traditional ML: SVM for Interface Prediction (PCPIP)

Representing traditional machine learning approaches, PCPIP uses a Support Vector Machine (SVM) to classify protein-protein interaction (PPI) interfaces as native or non-native, which is valuable for validating interactions between R-proteins and pathogen effectors [42].

Methodology:

• Feature Engineering: The classifier relies on manually curated interface properties calculated by the PISA software, including accessible surface area (ASA), buried surface area (BSA), dissociation free energy, hydrogen bonds, and salt bridges [42].
• Training and Validation: The SVM model is trained on known dimer complexes and evaluated on benchmarking datasets, showing strong performance in distinguishing biologically relevant interfaces. This approach highlights the utility of expert-curated features, even alongside more complex DL models [42].

The workflow below illustrates the typical process for identifying and analyzing NBS-LRR genes, from initial identification to functional validation.

Diagram 1: R-gene Analysis Workflow

Successful R-gene prediction and analysis relies on a suite of computational tools and databases. The table below lists key resources.

Table 2: Essential Research Reagents and Resources

Category	Tool/Database	Primary Function
Genome Databases	NCBI Genome, Phytozome, Plaza, GDR	Source of plant genome sequences and annotations [3] [40]
Domain Analysis	HMMER, Pfam, NCBI-CDD, SMART	Identification of NBS, TIR, CC, LRR domains [3] [39] [40]
Evolutionary Analysis	OrthoFinder, MAFFT, IQ-TREE	Orthogroup clustering and phylogenetic tree construction [3]
Expression Analysis	IPF Database, CottonFGD, NCBI BioProject	RNA-seq data for expression profiling under stress [3]
Structure Prediction	AlphaFold2, P2Rank	Protein structure prediction and ligand-binding site analysis [43] [44]
Interaction Validation	PCPIP, STRING, BioGRID	PPI interface classification and known interaction data [45] [42]

Integration with Comparative Analysis of NBS Domain Genes

Machine learning classifiers are profoundly enhancing large-scale comparative genomic studies of NBS genes. For instance, a recent analysis of 12,820 NBS genes across 34 plant species identified 168 distinct domain architecture classes, revealing both classical and species-specific patterns [3]. Tools like PRGminer can rapidly and accurately annotate such vast datasets, enabling researchers to focus on evolutionary analysis. This study further utilized expression profiling to identify key orthogroups (OGs) upregulated in response to cotton leaf curl disease and employed virus-induced gene silencing (VIGS) to validate the role of a specific NBS gene (GaNBS in OG2) in viral defense [3]. The ability of DL models like DPFunc to pinpoint key functional residues [41] can directly inform such validation experiments by highlighting candidate regions for mutagenesis.

The integration of machine and deep learning classifiers into the plant immunology toolkit marks a significant advancement over traditional, homology-based methods for R-gene discovery. As demonstrated, tools like PRGminer offer high-throughput, accurate prediction and classification, while approaches like DPFunc provide deeper functional insights by linking sequence and structure to biological role. When used in conjunction with established evolutionary and expression analysis techniques, these classifiers empower researchers to decipher the complex landscape of plant disease resistance genes more efficiently and at an unprecedented scale, accelerating the development of resilient crop varieties.

Plant immunity relies on a sophisticated defense system where nucleotide-binding site (NBS) domain genes play a pivotal role as intracellular immune receptors. These genes, particularly those belonging to the NBS-LRR (NLR) family, constitute one of the largest and most variable gene families in plants, responsible for recognizing pathogen effector proteins and initiating robust immune responses [3] [46]. The NBS domain serves as the molecular switch that binds and hydrolyzes ATP, providing the energy for activating downstream defense signaling pathways [46]. Understanding the diversity, evolution, and function of these genes across plant species is crucial for developing disease-resistant crops, yet their extensive diversification presents significant research challenges.

Specialized bioinformatics databases have become indispensable tools for navigating the complexity of NLR genes. This guide provides a comparative analysis of three specialized resources—ANNA, PlaRRP, and DRAGO—focusing on their applications in comparative genomics and functional characterization of NBS domain genes. We evaluate their scope, data content, analytical capabilities, and utility for researchers aiming to identify novel resistance genes for crop improvement.

Database Comparison: Scope, Features, and Applications

The landscape of specialized databases for NBS gene research varies significantly in scope, data content, and functionality. The table below provides a systematic comparison of ANNA and DRAGO based on available information. Notably, comprehensive details for PlaRRP could not be ascertained from the search results.

Table 1: Comparative Analysis of Specialized Databases for NBS Gene Research

Feature	ANNA (Angiosperm NLR Atlas)	DRAGO (Disease Resistance Analysis and Gene Orthology)	PlaRRP
Primary Focus	Census and classification of NLR genes across angiosperms [3]	Annotation of resistance genes from sequence data [47]	Information not available in search results
Data Content	>90,000 NLR genes from 304 angiosperm genomes [3]	Not a pre-populated database; an analysis pipeline [47]	Information not available in search results
Key Utility	Evolutionary studies, comparative genomics, identifying lineage-specific expansions/losses [3]	Functional annotation of user-submitted sequences, domain architecture prediction [47]	Information not available in search results
Domain Detection	Implied from curated data	Hidden Markov Models (HMMs) for LRR, Kinase, NBS, TIR; COILS for CC; TMHMM for TM [47]	Information not available in search results
Access	Presumably a queryable database	Web interface (PRGdb) & API for large-scale analysis [47]	Information not available in search results

Analysis of Database Capabilities and Research Applications

• ANNA excels in providing evolutionary context for NLR genes. Its extensive curated data allows researchers to identify patterns of gene family expansion and contraction across angiosperms. For example, studies have used such data to note the complete loss of TNL genes in monocots like rice and the significant reduction in TNL and RNL subfamilies in certain eudicots like Salvia miltiorrhiza [46]. This makes ANNA ideal for generating evolutionary hypotheses and selecting candidate genes from diverse plant lineages.

• DRAGO functions as an analytical pipeline rather than a pre-populated database. Its strength lies in annotating custom sequence data (genomes or transcriptomes), making it invaluable for studying non-model organisms or newly sequenced species. DRAGO automatically detects key resistance gene domains and provides a standardized classification, which was a critical step in genome-wide studies identifying 196 NBS-LRR genes in Salvia miltiorrhiza and 239 in tung trees (Vernicia species) [46] [48]. Its API access facilitates the high-throughput analysis needed for large genomic datasets.

Experimental Workflows for Functional Characterization

Translating bioinformatic predictions from databases like ANNA and DRAGO into validated resistance genes requires a robust experimental pipeline. The following workflow, derived from recent literature, outlines the key steps from identification to functional validation of NBS-LRR genes.

Diagram 1: Functional Gene Validation Workflow

Detailed Experimental Protocols

Genome-Wide Identification and Domain Analysis

The initial phase involves the comprehensive identification of NBS-encoding genes within a target genome.

• Methodology: Employ Hidden Markov Model (HMM) searches against the proteome of the target species. The standard protocol uses HMMER software with the Pfam database (e.g., PF00931 for NB-ARC domain) and a stringent e-value cutoff (e.g., 1.1e-50) to minimize false positives [3] [48]. Subsequent domain architecture analysis (for CC, TIR, LRR domains) is performed using complementary tools like COILS and TMHMM [47].
• Application: This methodology identified 12,820 NBS-domain-containing genes across 34 plant species in a large-scale comparative study [3]. In another example, this approach revealed 196 NBS genes in the medicinal plant Salvia miltiorrhiza, of which only 62 possessed complete N-terminal and LRR domains to be classified as typical NLRs [46].

Transcriptional Profiling and Candidate Selection

• Methodology: Utilize RNA-seq data from various tissues and stress conditions (biotic and abiotic) to generate expression profiles (e.g., in FPKM or TPM units). Analyze data to identify NLRs with high constitutive expression, as recent studies show that functional NLRs are often highly expressed in uninfected plants [49]. For example, in Arabidopsis thaliana, known functional NLRs are significantly enriched in the top 15% of expressed NLR transcripts [49].
• Application: This principle was successfully applied in wheat, where researchers generated a transgenic array of 995 NLRs selected based on high expression signatures. This led to the identification of 31 new resistance genes (19 against stem rust and 12 against leaf rust) [49].

Functional Validation via Virus-Induced Gene Silencing (VIGS)

The final step confirms the biological function of candidate NBS-LRR genes in plant immunity.

• Protocol:
  • Vector Construction: Clone a 150-300 bp fragment of the candidate NBS-LRR gene into a VIGS vector (e.g., derived from Tobacco Rattle Virus).
  • Plant Transformation: Introduce the vector into plants of interest via Agrobacterium tumefaciens-mediated infiltration.
  • Pathogen Challenge: Inoculate silenced plants with the target pathogen and monitor for disease symptoms over time.
  • Molecular Confirmation: Use qRT-PCR to verify the reduction of target gene expression in silenced plants.
• Application: This method confirmed the role of the GaNBS (OG2) gene in resistant cotton, where its silencing led to increased viral titer [3]. Similarly, silencing of Vm019719 in the resistant tung tree Vernicia montana compromised its resistance to Fusarium wilt, confirming the gene's essential role in immunity [48].

Essential Research Reagents and Solutions

Successful research in this field relies on a suite of specialized reagents and computational tools. The following table catalogues key resources for the experimental and bioinformatic workflows described.

Table 2: Key Research Reagent Solutions for NBS Gene Analysis

Reagent / Tool	Function / Application	Specifications / Examples
HMMER Software	Identifies protein domains (NBS, LRR, TIR) in sequence data [3] [48]	Used with Pfam HMM profiles (e.g., NB-ARC PF00931) [3]
COILS / TMHMM	Predicts coiled-coil (CC) and transmembrane (TM) domains [47]	Integrated into the DRAGO pipeline for domain annotation [47]
OrthoFinder	Determines orthogroups and gene families across species [3]	Used for evolutionary analysis; identified 603 NBS orthogroups [3]
VIGS Vectors	Functional validation through transient gene silencing [3] [48]	TRV-based vectors for Agrobacterium-mediated delivery [48]
RNA-seq Datasets	Expression profiling under biotic/abiotic stress [3] [49]	Publicly available in NCBI SRA, IPF, and species-specific databases [3]

The comparative analysis of specialized databases reveals complementary strengths. ANNA provides an unparalleled evolutionary resource for exploring the macro-evolution of NLRs across angiosperms, while DRAGO offers a flexible, powerful pipeline for annotating resistance genes in novel sequence data. The inability to profile PlaRRP here highlights the dynamic nature of bioinformatics resources and the need for researchers to consult the most current literature.

The integration of these bioinformatic resources with standardized experimental workflows—from genome-wide identification and expression analysis to functional validation via VIGS—creates a powerful pipeline for accelerating the discovery of new R genes. This integrated approach, leveraging both computational and experimental tools, is already yielding tangible results, identifying new sources of resistance against devastating diseases in staple crops, and holds great promise for future disease-resistance breeding programs [49].

Genome-Wide Identification and Orthogroup Analysis

Genome-wide identification and orthogroup analysis represent foundational methodologies in modern comparative genomics, enabling researchers to decipher gene family evolution, functional diversification, and adaptive processes across species. These approaches are particularly valuable for studying large and complex gene families involved in critical biological processes, such as plant immunity. The nucleotide-binding site (NBS) domain gene family, which encompasses key plant disease resistance genes (NLRs), exemplifies a system where these methods have revealed remarkable evolutionary dynamics and functional specialization [3]. This guide objectively compares experimental approaches for genome-wide identification and orthogroup analysis of such gene families, focusing specifically on NBS domain genes across plant species, and provides researchers with standardized frameworks for implementing these analyses in their systems.

Core Principles and Definitions

Genome-wide identification refers to the comprehensive cataloging and characterization of all members of a specific gene family within a fully sequenced genome. This process typically involves domain-based searches, phylogenetic reconstruction, and structural analysis [3] [50] [51].

Orthogroup analysis clusters genes into families descended from a single gene in the last common ancestor of all species being compared. OrthoFinder is the most widely used tool for this purpose, employing a graph-based algorithm to infer orthogroups from sequence similarity data [3] [50] [52]. This method objectively circumscribes gene families across multiple species, enabling systematic comparative analyses.

For NBS domain genes specifically, classification systems have been established based on domain architecture, including:

• CNL: Coiled-coil-NBS-LRR
• TNL: TIR-NBS-LRR
• RNL: RPW8-NBS-LRR
• NL: NBS-LRR (without distinctive N-terminal domain)
• N: NBS-only variants [3] [50] [51]

Comparative Methodological Approaches

Genome-Wide Identification Protocols

Table 1: Comparative Methodologies for NBS Gene Identification Across Plant Species

Methodological Step	Hussain et al. (2024) 34 Species [3]	Asparagus Study (2025) 3 Species [50]	Nicotiana Study (2025) 3 Species [51]
Domain Identification	PfamScan.pl HMM with NB-ARC domain (PF00931), e-value 1.1e-50	HMMER with PF00931, CDD validation	HMMER v3.1b2 with PF00931, additional TIR/LRR domains
Classification System	168 architecture classes, species-specific patterns	TNL/CNL/RNL with truncations	8 subfamilies based on domain composition
Validation Approach	Domain architecture consistency	InterProScan, NCBI CD-Search	NCBI CDD, protein completeness check
Genomic Distribution	Tandem duplication analysis	Chromosomal clustering (≤8 gene spacing)	MCScanX for duplication patterns

Orthogroup Analysis Frameworks

Table 2: Orthogroup Analysis Methods and Outcomes

Analysis Component	Hussain et al. (2024) [3]	Asparagus Comparative Study [50]	Tool-Based Solutions
Primary Software	OrthoFinder v2.5.1	OrthoFinder v2.2.7	PlantTribes2 Galaxy implementation
Sequence Search	DIAMOND for fast similarity	BLAST-based bit score normalization	BLAST, HMMER options
Clustering Method	MCL algorithm	OrthoFinder default	Customizable algorithms
Orthogroup Output	603 orthogroups across 34 species	16 conserved NLR pairs between A. setaceus and A. officinalis	Pre-computed orthologous families
Core Orthogroups	OG0, OG1, OG2 as most common	Species-specific conservation patterns	Core orthogroup (CROG) analysis

Experimental Data from Comparative Studies

Quantitative Findings in Plant Species

Table 3: NBS Gene Distribution Across Plant Taxa

Plant Species/Group	Total NBS Genes	CNL Components	TNL Components	Other Architectures	Study
34 Land Plants	12,820 genes	Not specified	Not specified	168 domain architecture classes	[3]
Nicotiana tabacum	603	23.3% CC-NBS	2.5% TIR-NBS	45.5% NBS-only	[51]
Nicotiana sylvestris	344	Similar distribution	Similar distribution	Similar distribution	[51]
Nicotiana tomentosiformis	279	Similar distribution	Similar distribution	Similar distribution	[51]
Asparagus setaceus	63	Classified by TNL/CNL/RNL	Classified by TNL/CNL/RNL	Classified by TNL/CNL/RNL	[50]
Asparagus kiusianus	47	Classified by TNL/CNL/RNL	Classified by TNL/CNL/RNL	Classified by TNL/CNL/RNL	[50]
Asparagus officinalis	27	Classified by TNL/CNL/RNL	Classified by TNL/CNL/RNL	Classified by TNL/CNL/RNL	[50]

Expression and Functional Validation

Comparative expression profiling in cotton orthogroups demonstrated that OG2, OG6, and OG15 showed upregulated expression in various tissues under biotic and abiotic stresses [3]. Functional validation through virus-induced gene silencing (VIGS) of GaNBS (OG2) in resistant cotton demonstrated its putative role in virus tittering, confirming the functional significance of conserved orthogroups [3].

In Asparagus species, pathogen inoculation assays revealed distinct phenotypic responses: domesticated A. officinalis was susceptible to Phomopsis asparagi, while wild A. setaceus remained asymptomatic. Notably, most preserved NLR genes in A. officinalis showed either unchanged or downregulated expression following fungal challenge, indicating potential functional impairment during domestication [50].

Standardized Experimental Workflows

Genome-Wide Identification Protocol

Figure 1: Workflow for genome-wide identification of NBS domain genes. The process begins with data acquisition and proceeds through domain searches and classification.

Orthogroup Analysis Pipeline

Figure 2: Orthogroup analysis pipeline from sequence input to functional validation.

Research Reagent Solutions

Table 4: Essential Research Reagents and Computational Tools

Reagent/Tool	Specific Function	Application in NBS Studies
HMMER Suite	Hidden Markov Model searches	NB-ARC domain (PF00931) identification	[3] [50] [51]
OrthoFinder	Orthogroup inference from sequences	Pan-species orthogroup clustering	[3] [50] [52]
DIAMOND	Accelerated BLAST-compatible search	Fast sequence similarity for large datasets	[3]
MCScanX	Genome collinearity and duplication	Tandem and segmental duplication analysis	[50] [51]
PlantTribes2	Gene family classification framework	Scalable analysis in Galaxy environment	[52]
MEME Suite	Motif discovery and analysis	Conserved motif identification in NBS domains	[50]
InterProScan	Protein domain classification	Multi-domain architecture validation	[50]

Technical Considerations and Best Practices

Genome Assembly Quality Dependence

The quality of genome-wide identification is heavily dependent on the completeness and continuity of genome assemblies. While short-read assemblies (e.g., Illumina) might capture most coding regions, they are often fragmented and poorly resolve repetitive elements [53] [54]. For large, complex gene families like NBS genes, long-read sequencing technologies (PacBio HiFi, ONT) combined with chromatin conformation capture (Hi-C) provide chromosome-scale assemblies that enable more comprehensive identification and accurate genomic distribution analysis [53] [54].

Taxonomic Sampling Strategy

Comparative studies across multiple species consistently demonstrate that taxonomic sampling strategy significantly impacts orthogroup inference. The 34-species analysis revealed both core orthogroups (e.g., OG0, OG1, OG2) present across most species and unique orthogroups specific to particular lineages [3]. Including evolutionarily diverse representatives from different clades enables more accurate reconstruction of gene family evolutionary history.

Validation and Quality Control

Rigorous validation steps are essential for both genome-wide identification and orthogroup analysis. For NBS gene identification, this includes:

• Verification of conserved NBS motifs (P-loop, GLPL, MHD, Kinase 2) [50]
• Manual inspection of domain architecture completeness
• Assessment of gene models using transcriptomic evidence [3] [50]

For orthogroup analysis, quality control measures include:

• Examination of orthogroup size distributions
• Assessment of species-specific gene loss patterns
• Validation with known gene families [3] [52]

Genome-wide identification and orthogroup analysis provide powerful complementary approaches for understanding the evolution and functional diversification of gene families across plant species. Standardized methodologies employing HMMER, OrthoFinder, and complementary bioinformatic tools have enabled robust comparative analyses of NBS domain genes, revealing significant expansion and contraction patterns, lineage-specific adaptations, and conserved orthogroups with potential functional significance. The continued refinement of these methodologies, coupled with improving genome assembly quality and expanding taxonomic sampling, will further enhance our understanding of plant immune gene evolution and facilitate the development of disease-resistant crop varieties through informed breeding strategies.

Expression Profiling Using RNA-seq and Transcriptomic Data

The nucleotide-binding site (NBS) domain genes represent one of the largest and most critical families of plant resistance (R) genes, playing indispensable roles in effector-triggered immunity (ETI) by encoding intracellular proteins capable of recognizing pathogen-derived effectors and activating robust defense responses [3] [46]. These genes typically feature a conserved NBS domain alongside leucine-rich repeat (LRR) domains and are classified into distinct subfamilies—TIR-NBS-LRR (TNL), CC-NBS-LRR (CNL), and RPW8-NBS-LRR (RNL)—based on their N-terminal domains [9] [46]. Expression profiling using RNA sequencing (RNA-seq) and transcriptomic analyses has emerged as a powerful approach for investigating the roles of NBS domain genes in plant-pathogen interactions, identifying candidate resistance genes, and understanding the molecular basis of immune responses across diverse plant species [55] [3]. This guide provides a comparative analysis of experimental approaches, data interpretation strategies, and methodological considerations for employing RNA-seq in the study of NBS-mediated resistance, with supporting experimental data from recent research.

Comparative Analysis of RNA-seq Applications in NBS Gene Research

Experimental Designs and Key Findings

RNA-seq technology has been deployed across various plant species to investigate NBS domain gene expression under different experimental conditions, particularly in response to pathogen challenge. The table below summarizes key studies, their experimental designs, and principal findings related to NBS gene expression.

Table 1: Comparative summary of RNA-seq studies on NBS domain genes in plant-pathogen interactions

Plant Species	Experimental Design	Key Findings Related to NBS Genes	Reference
Gossypium hirsutum (Upland cotton)	Two NILs with/without Renbarb2 QTL; inoculated with reniform nematode; RNA-seq at 5-, 9-, 13 days after inoculation	Identified 966 DEGs in resistant NIL vs. 133 in susceptible; Gohir.D11G302300 (CC-NBS-LRR) showed ~3.5-fold higher basal expression in resistant roots	[55]
Multiple species (34 total)	Comparative analysis of 12,820 NBS domain genes across species; expression profiling under biotic/abiotic stresses	Identified 603 orthogroups; OG2, OG6, OG15 upregulated in tolerant genotypes under cotton leaf curl disease; 6583 unique variants in tolerant cotton vs. 5173 in susceptible	[3]
Brassica oleracea (Cabbage)	RNA-seq of cabbage challenged with Fusarium oxysporum; digital gene expression and RT-PCR validation	14 TNL genes responded significantly to Fusarium infection; 9 upregulated and 5 downregulated; Foc1 works with clustered genes for resistance	[56]
Salvia miltiorrhiza (Medicinal plant)	Genome-wide identification of 196 NBS-LRR genes; transcriptome analysis under stress conditions	62 typical NLRs identified; expression closely associated with secondary metabolism; promoter analysis revealed hormone and stress-responsive elements	[46] [57]
Wheat and Barley	Comparative population genomics across 672 wheat and 679 barley accessions; exome sequencing	Identified 451 orthogroups under convergent selection; homeolog-specific selection patterns in polyploid wheats	[58]

Insights from Comparative Studies

The comparative analysis of NBS domain genes across multiple plant species reveals several key patterns. First, resistant genotypes typically activate a broader and more sustained defense transcriptome, as evidenced by the identification of 966 differentially expressed genes (DEGs) in resistant cotton NILs compared to only 133 DEGs in susceptible lines following nematode infection [55]. Second, specific NBS gene subfamilies demonstrate distinct evolutionary patterns across plant lineages, with TNL genes being abundant in dicots but absent in monocots, while some medicinal plants like Salvia miltiorrhiza show marked reduction in TNL and RNL subfamilies [46]. Third, expression profiling successfully identifies candidate resistance genes through combined analysis of differential expression, genetic variation, and genomic position within known quantitative trait loci (QTL) [55] [3].

Experimental Protocols for RNA-seq in NBS Gene Research

Standard Workflow for Transcriptome Profiling

The following diagram illustrates the comprehensive workflow for RNA-seq-based expression profiling of NBS domain genes, integrating both standard transcriptomic approaches and specialized analyses for resistance gene studies:

Detailed Methodological Considerations

Experimental Design and Sample Preparation

Robust experimental design forms the foundation for reliable RNA-seq studies of NBS genes. Research on cotton nematode resistance employed nearly isogenic lines (NILs) differing only at the target resistance locus, allowing researchers to attribute expression differences specifically to the resistance QTL rather than background genetic variation [55]. The time-course design with samples collected at 5, 9, and 13 days after inoculation enabled capturing both early and late defense responses. Studies typically pool root systems from multiple plants (typically 3) to constitute a single biological replicate, with three or more independent biological replicates per condition to ensure statistical robustness [55]. For RNA extraction, the sodium hypochlorite washing method is commonly employed to remove nematodes from root tissues before RNA extraction, ensuring that transcriptomic data represents plant responses rather than a mixture of plant and pathogen transcripts [55].

Library Preparation and Sequencing

Standard RNA-seq protocols begin with RNA quality assessment using instruments like Bioanalyzer to ensure RNA Integrity Numbers (RIN) exceed 8.0. Most studies employ poly-A selection for mRNA enrichment followed by cDNA library preparation using kits such as Illumina TruSeq Stranded mRNA. Sequencing is typically performed on Illumina platforms (HiSeq or NovaSeq) to generate 100-150 bp paired-end reads, with recommended sequencing depth of 20-40 million reads per sample to ensure sufficient coverage for both highly and lowly expressed transcripts [55] [56].

Bioinformatics Analysis Pipeline

The bioinformatics workflow for NBS gene expression studies involves several critical steps. After quality control with FastQC and adapter trimming, reads are aligned to a reference genome using splice-aware aligners like STAR or HISAT2 [55] [3]. For non-model species without reference genomes, de novo transcriptome assembly using tools like Trinity may be employed. Gene expression is quantified using featureCounts or HTSeq, followed by normalization using FPKM or TPM values to enable cross-sample comparison [3]. Differential expression analysis is typically performed using DESeq2 or edgeR, applying appropriate multiple testing correction (Benjamini-Hochberg FDR < 0.05) [55] [56].

Specialized Analysis for NBS Domain Genes

NBS domain gene identification represents a specialized component of the analysis pipeline. Researchers use Hidden Markov Model (HMM) searches with the Pfam NBS (NB-ARC) domain (PF00931) against the protein sequences of target species, typically applying an E-value cutoff of 1e-10 [3] [9] [46]. Additional domain analysis using Pfam, SMART, and Paircoil2 helps classify NBS genes into subfamilies (TNL, CNL, RNL) based on N-terminal domains [56] [46]. For cross-species comparisons, orthogroup analysis using OrthoFinder identifies evolutionarily conserved NBS gene groups, enabling researchers to track expression patterns of orthologous genes across different species [3] [9]. Integration of SNP data from RNA-seq reads enables identification of non-synonymous mutations in NBS genes that may contribute to functional differences between resistant and susceptible genotypes [55] [3].

Table 2: Key research reagents, databases, and computational tools for RNA-seq studies of NBS domain genes

Category	Item/Resource	Function/Application	Examples/Specifications
Wet Lab Reagents	RNA extraction kits	High-quality RNA isolation from plant tissues	TRIzol, RNeasy Plant Mini Kit
	Library preparation kits	cDNA library construction for sequencing	Illumina TruSeq Stranded mRNA
	qRT-PCR reagents	Validation of RNA-seq results	SYBR Green, TaqMan assays
Bioinformatics Tools	Quality control tools	Assessment of raw sequence data quality	FastQC, MultiQC
	Alignment software	Mapping reads to reference genomes	STAR, HISAT2, Bowtie2
	Assembly programs	De novo transcriptome assembly	Trinity, SOAPdenovo-Trans
	Differential expression	Identifying significantly changed genes	DESeq2, edgeR, limma-voom
	NBS identification	HMM-based NBS domain detection	HMMER, PfamScan
	Orthogroup analysis	Cross-species gene family analysis	OrthoFinder, InParanoid
Databases	Genome databases	Reference genomes and annotations	Phytozome, NCBI, Ensembl Plants
	Domain databases	Protein domain identification and classification	Pfam, SMART, InterPro
	Expression databases	Repository for transcriptome data	IPF database, CottonFGD, NCBI GEO
Validation Approaches	VIGS systems	Functional validation of candidate NBS genes	Tobacco rattle virus (TRV)-based vectors
	CRISPR-Cas9	Targeted gene knockout for functional studies	Plasmid constructs, ribonucleoproteins

Data Interpretation and Integration Strategies

Signaling Pathways in NBS-Mediated Resistance

RNA-seq studies have elucidated key signaling pathways activated during NBS-mediated resistance responses. The following diagram illustrates the integrated signaling network based on transcriptomic analyses of resistant plants:

Key Expression Patterns and Regulatory Networks

Transcriptomic analyses consistently identify several hallmark expression patterns associated with effective NBS-mediated resistance. Resistant genotypes typically exhibit sustained upregulation of defense-related genes across multiple timepoints, whereas susceptible plants show only transient or minimal induction [55]. The redox homeostasis and oxidation-reduction processes are commonly enriched in resistant plants, with genes involved in these pathways being upregulated at early infection stages (5-9 days after inoculation) [55]. Transcription factor families including ERF, WRKY, and NAC show pronounced enrichment in resistant genotypes, particularly at later timepoints (13 days after inoculation), suggesting their importance in maintaining defense responses [55]. Additionally, cell wall reinforcement genes are typically upregulated during early infection stages in resistant plants, contributing to physical barriers against pathogen penetration [55].

Expression profiling using RNA-seq and transcriptomic data has revolutionized our understanding of NBS domain gene function, evolution, and regulation in plant immunity. The comparative analysis presented in this guide demonstrates both conserved and species-specific aspects of NBS gene expression across diverse plant-pathogen systems. Future directions in this field will likely include single-cell transcriptomic approaches to resolve NBS gene expression at cellular resolution [59], integration of multi-omics data for comprehensive understanding of resistance mechanisms [60], and machine learning applications for predicting resistance function from sequence and expression features. The continued refinement of RNA-seq methodologies and analytical frameworks will further enhance our ability to identify and characterize NBS domain genes, ultimately accelerating the development of disease-resistant crop varieties through molecular breeding.

Overcoming Challenges in NBS Gene Family Analysis and Interpretation

Addressing Gene Loss and Domain Architecture Variation

The nucleotide-binding site (NBS) domain represents a fundamental component of the plant immune system, serving as the central signaling module in a vast family of disease resistance (R) genes. These genes, typically encoding proteins with NBS and leucine-rich repeat (LRR) domains, constitute the largest and most prominent class of R genes in plants, enabling recognition of diverse pathogens including viruses, bacteria, fungi, and oomycetes [61] [62]. The evolution of this gene family is characterized by remarkable dynamism, driven by gene duplication, diversifying selection, and frequent gene loss events, which collectively shape the specific resistance repertoire of each plant species [20] [61]. This guide provides a comparative analysis of NBS domain genes across plant species, focusing on the phenomena of gene loss and domain architecture variation. We objectively compare the performance of different methodologies for studying these genes and present supporting experimental data, including recent findings from 2024. The insights are framed within the broader thesis that understanding this genomic plasticity is crucial for deciphering plant adaptation mechanisms and for engineering durable disease resistance in crops.

Classification and Genomic Distribution of NBS Domain Architectures

Major Domain Architectures and Their Prevalence

The core structure of an NBS-LRR protein includes an N-terminal domain, a central NBS (or NB-ARC) domain, and a C-terminal LRR domain [61]. The N-terminal domain is a primary source of variation and defines two major subfamilies: the TIR-NBS-LRR (TNL) proteins, which contain a Toll/Interleukin-1 Receptor domain, and the CC-NBS-LRR (CNL) proteins, which typically possess a Coiled-Coil domain [3] [61]. A third, smaller subclass features an RPW8 domain at the N-terminus and is designated RNL [6]. Beyond these canonical architectures, numerous other domain combinations exist, including truncated forms that lack one or more of the core domains (e.g., TIR-NBS, CC-NBS, NBS-LRR, and NBS-only proteins) [3] [5].

Table 1: Diversity of NBS Domain Architectures Across Plant Species

Plant Species	Total NBS Genes	TNL	CNL	RNL	Truncated/Other	Key References
Arabidopsis thaliana	~150	~62	~85	~3	~58 (TN, CN, etc.)	[61]
Vernicia montana (Resistant Tung)	149	3	9	-	137 (CC-NBS, TIR-NBS, NBS, etc.)	[5]
Vernicia fordii (Susceptible Tung)	90	0	12	-	78 (CC-NBS, NBS, etc.)	[5]
Akebia trifoliata	73	19	50	4	-	[6]
Common Bean (Phaseolus vulgaris)	323 (178 full + 145 partial)	30	148	-	-	[62]
Cassava (Manihot esculenta)	327 (228 full + 99 partial)	34	128	-	-	[63]

Genomic Organization and Clustering

A hallmark of NBS-encoding genes is their tendency to be organized in clusters within plant genomes. This clustering, resulting from both segmental and tandem duplications, is a key genomic feature that facilitates the rapid evolution of new resistance specificities through unequal crossing-over and gene conversion [63] [61]. For instance, in cassava, 63% of the 327 identified R genes are arranged in 39 clusters across the chromosomes, most of which are homogeneous (containing genes from a recent common ancestor) [63]. Similarly, in Akebia trifoliata, 41 of its 73 NBS genes are located in clusters, predominantly at chromosome ends, while the remaining 23 are singletons [6]. This clustered organization stands in stark contrast to the more uniform distribution of most other plant genes and underscores the evolutionary battle between plants and their rapidly evolving pathogens.

Evidence and Impact of Gene Loss and Gain

The birth-and-death evolution model, involving repeated gene duplication and loss, is a dominant force shaping NBS-LRR repertoires [20] [61]. Lineage-specific gene loss is particularly evident in the distribution of the TNL class. A striking example is the complete absence of TNL genes in monocot cereals, suggesting a loss in the ancestral cereal lineage [61]. This loss is not confined to monocots; several eudicot species also show a reduction or complete lack of TNLs. For example, the susceptible tung tree (Vernicia fordii) has completely lost TNL genes, whereas its resistant counterpart (Vernicia montana) retains 12 TNLs, indicating a potential correlation between gene loss and susceptibility [5]. Similar losses have been reported in sesame (Sesamum indicum) [5].

Gene loss is not limited to entire classes. The loss of specific LRR domains can also impact resistance. In Vernicia species, the LRR1 and LRR4 domains were found exclusively in the resistant V. montana and were absent in the susceptible V. fordii, suggesting that the loss of these specific domains may have compromised the immune recognition capacity of V. fordii [5]. Conversely, gene gain through duplication is a critical mechanism for expanding the resistance arsenal. In A. trifoliata, tandem and dispersed duplications were identified as the main forces for NBS expansion, responsible for 33 and 29 genes, respectively [6].

Table 2: Documented Cases of Gene and Domain Loss in NBS Genes

Plant Species	Type of Loss	Genomic Consequence	Putative Phenotypic Impact	Citation
Cereals (e.g., Rice, Maize)	Loss of entire TNL class	Complete absence of TNL-type NBS genes	Altered signaling pathways; reliance on CNL-type genes	[61]
Vernicia fordii (Tung Tree)	Loss of entire TNL class; Loss of LRR1, LRR4 domains	90 NBS genes vs. 149 in resistant relative; Reduced LRR diversity	Susceptibility to Fusarium wilt	[5]
Sesamum indicum (Sesame)	Loss of entire TNL class	Absence of TIR-domain containing NBS-LRRs	Not specified	[5]

Methodologies for Comparative Analysis and Functional Validation

Genomic Identification and Classification Pipelines

A standard, robust pipeline for identifying NBS genes from genome sequences relies on homology searches using Hidden Markov Models (HMMs). The typical workflow begins by scanning predicted protein sequences from a genome assembly with a Pfam HMM profile for the NB-ARC domain (PF00931) [3] [63] [6]. Candidate genes are then filtered based on E-value significance (e.g., < 1x10⁻²⁰). To improve sensitivity, a species-specific HMM can be built from the initial high-quality candidates and used to re-search the proteome [63]. Subsequent domain architecture classification is performed using additional HMM profiles (e.g., for TIR, RPW8, LRR domains) and tools like Coiled-coil prediction algorithms, as CC domains are not always detected by Pfam alone [63] [6]. Manual curation is essential to remove false positives, such as proteins with kinase domains that share minimal similarity with NBS domains [63].

Orthogroup and Evolutionary Analysis

To understand deep evolutionary relationships and gene loss events, orthogroup analysis across multiple species is powerful. Tools like OrthoFinder are used to cluster NBS protein sequences from a wide range of plant species into orthogroups (OGs)—groups of genes descended from a single gene in the last common ancestor [3]. This analysis can reveal core orthogroups (common across many species) and unique or lineage-specific orthogroups. For example, a 2024 study identified 603 orthogroups from 34 plant species, with some core OGs (e.g., OG0, OG1, OG2) being widely represented, while others (e.g., OG80, OG82) were highly specific to certain species, indicative of lineage-specific gene retention or loss [3].

Functional Validation Using Virus-Induced Gene Silencing (VIGS)

The ultimate test of an NBS gene's function is experimental validation. Virus-Induced Gene Silencing (VIGS) is a rapid, powerful technique used to transiently knock down the expression of a candidate gene in a plant and assess the resulting change in phenotype, typically disease resistance.

Detailed VIGS Protocol (as applied in recent studies):

• Candidate Gene Selection: Select a target NBS gene based on expression data (e.g., upregulation in resistant varieties upon infection) or genetic association [3] [5].
• Vector Construction: A unique, gene-specific fragment (typically 200-500 bp) of the candidate gene is cloned into a VIGS vector (e.g., based on Tobacco Rattle Virus, TRV).
• Plant Inoculation: The recombinant VIGS vector is introduced into plants. For Agrobacterium tumefaciens-mediated delivery, the vector is transformed into Agrobacterium cells, which are then infiltrated into the leaves of young plants. For tung trees, researchers used cotyledons from 3-day-old seedlings [5].
• Phenotypic Assessment: After a period allowing for gene silencing and viral spread, plants are challenged with the target pathogen. The disease symptoms and pathogen biomass are compared between plants silenced for the candidate gene and control plants (e.g., silenced with an empty vector).
• Molecular Confirmation: The silencing efficiency is confirmed using quantitative reverse transcription PCR (qRT-PCR) to measure the transcript levels of the target gene.

A compelling application of VIGS demonstrated that silencing the GaNBS gene (from orthogroup OG2) in resistant cotton led to a significant increase in viral titer, confirming its putative role in resistance to cotton leaf curl disease [3]. Similarly, in the resistant tung tree V. montana, VIGS of the Vm019719 (a NBS-LRR gene) compromised its resistance to Fusarium wilt, providing direct evidence of its function [5].

Table 3: Essential Reagents and Resources for NBS Gene Research

Resource Category	Specific Tool / Database / Reagent	Function and Application	Example
Genomic Databases	Plant GARDEN, Phytozome, Ensembl Plants	Access to assembled plant genomes, gene annotations, and comparative genomics tools.	Plant GARDEN provides 304 assembled genomes from 234 species for cross-species analysis [64].
Domain Databases	Pfam, NCBI Conserved Domain Database (CDD)	Identify and annotate protein domains (NBS, TIR, LRR, CC) in candidate genes.	Pfam profile PF00931 (NB-ARC) is the standard for NBS domain identification [63] [6].
Analysis Software	HMMER, OrthoFinder, MEME Suite, MEGA	Perform sequence searches, orthogroup clustering, motif discovery, and phylogenetic analysis.	HMMER is used for initial HMM searches; OrthoFinder clusters genes into ortholog groups [3] [63].
Experimental Vectors	Virus-Induced Gene Silencing (VIGS) Vectors (e.g., TRV-based)	Transiently knock down gene expression in planta for functional validation.	Used to validate the role of GaNBS and Vm019719 in disease resistance [3] [5].
Specialized Databases	ANNA: An Angiosperm NLR Atlas, SolariX	Curated collections of NBS-LRR genes from hundreds of species or specific crops.	SolariX database compiles NBS-domain sequences from 96 potato cultivars [65].

Navigating Tandem Duplication Clusters and Sequence Divergence

The nucleotide-binding site (NBS) domain genes represent a critical superfamily of plant resistance (R) genes that function as intracellular immune receptors, enabling plants to recognize and respond to diverse pathogens [3] [32]. These genes, often characterized by their canonical NBS-leucine rich repeat (LRR) architecture, exhibit remarkable structural diversity and evolutionary dynamics across plant species [66]. Among the various mechanisms driving their evolution, tandem duplication stands out as a primary force for generating novel resistance specificities and adapting to rapidly evolving pathogens [67]. This process creates clusters of genetically similar NBS genes positioned in close proximity on chromosomes, serving as reservoirs for genetic innovation in plant immunity systems.

Understanding the principles governing tandem duplication clusters and sequence divergence is not merely an academic pursuit but a practical necessity for strategic crop improvement. Recent studies have demonstrated that the lineage-specific expansion of these gene clusters through tandem duplication significantly contributes to genotypic diversity and environmental adaptation in various plant species [67]. This comparative guide synthesizes experimental data and analytical methodologies to objectively evaluate how different plant species have navigated the complex landscape of NBS gene evolution, providing researchers with frameworks for identifying and harnessing these genetic elements for disease resistance breeding.

Comparative Genomic Landscape of NBS Domain Genes

Diversity in Gene Repertoire and Architecture

The genomic repertoire of NBS domain genes exhibits striking variation across plant species, reflecting distinct evolutionary paths and adaptation strategies. A comprehensive analysis of 34 plant species spanning from mosses to monocots and dicots identified 12,820 NBS-domain-containing genes, classified into 168 distinct architectural classes encompassing both classical and species-specific structural patterns [3]. This diversity includes not only well-characterized configurations like NBS, NBS-LRR, TIR-NBS, and TIR-NBS-LRR but also novel domain architectures such as TIR-NBS-TIR-Cupin1-Cupin1 and TIR-NBS-Prenyltransf that likely represent lineage-specific innovations [3].

The distribution of NBS genes across major crop and model plants reveals significant variation independent of genome size, with Fabaceae species exhibiting particular diversity in their NLR protein repertoire [68]. Pepper (Capsicum annuum L.) genomes harbor 252 NBS-LRR resistance genes, with an uneven distribution across all chromosomes and 54% forming 47 distinct gene clusters—a genomic organization primarily driven by tandem duplications and genomic rearrangements [66]. Similarly, medicinal plants like Salvia miltiorrhiza possess 196 NBS-LRR genes, though only 62 contain complete N-terminal and LRR domains, indicating substantial structural diversity within the gene family [46] [57].

Table 1: Comparative Analysis of NBS-LRR Genes Across Plant Species

Plant Species	Total NBS Genes	Clustered Genes	Major Subfamilies	Tandem Duplication Impact
Capsicum annuum (Pepper)	252	54% (47 clusters)	nTNL (248), TNL (4)	Dominant evolutionary mechanism [66]
Salvia miltiorrhiza (Danshen)	196	Not specified	CNL (61), RNL (1)	Marked reduction in TNL/RNL subfamilies [46]
Solanum tuberosum (Potato)	447	Lineage-specific clusters	CNL, TNL	Generates lineage-specific gene families [67]
Asparagus officinalis (Garden Asparagus)	27	Not specified	CNL, TNL, RNL	Contraction during domestication [19]
Arabidopsis thaliana	207	Not specified	CNL, TNL, RNL	Reference for comparative studies [46]

Subfamily Distribution and Evolutionary Patterns

The NBS-LRR gene family is primarily divided into three major subfamilies based on N-terminal domains: TIR-NBS-LRR (TNL), CC-NBS-LRR (CNL), and RPW8-NBS-LRR (RNL) [19]. Different plant lineages exhibit striking variations in subfamily representation, reflecting distinct evolutionary trajectories. Comparative analyses reveal a marked reduction in TNL and RNL subfamily members within Salvia species, with only 2 TNL and 1 RNL proteins identified among typical NLRs [46]. This pattern extends to monocotyledonous species such as Oryza sativa, Triticum aestivum, and Zea mays, where typical TNL and RNL subfamilies have been completely lost [57].

In contrast, gymnosperms like Pinus taeda exhibit significant expansion of the TNL subfamily, which comprises 89.3% of typical NBS-LRRs in this species [46]. The pepper genome demonstrates phylogenetic dominance of the nTNL subfamily over the TNL subfamily, with only 4 TNL genes identified among 252 NBS-LRR resistance genes [66]. These distribution patterns reflect deep evolutionary histories, with tandem duplication playing a crucial role in the lineage-specific expansion and contraction of these subfamilies across the plant kingdom.

Methodological Framework: Analyzing Tandem Duplication and Divergence

Genomic Identification and Classification

The accurate identification and classification of NBS domain genes form the foundation for comparative analysis. Current methodologies employ a dual approach combining Hidden Markov Model (HMM) searches and BLAST-based analyses to ensure comprehensive gene discovery [19]. The standard protocol begins with HMM searches using the conserved NB-ARC domain (Pfam: PF00931) as query, followed by local BLASTp analyses against reference NLR protein sequences with stringent E-value cutoffs (typically 1e-10) [19]. Candidate sequences identified through both methods are subsequently validated through domain architecture analysis using tools like InterProScan and NCBI's Batch CD-Search [19].

Following identification, genes are classified based on domain architecture into categories such as N (NBS only), NL (NBS-LRR), CN (CC-NBS), TN (TIR-NBS), CNL (CC-NBS-LRR), and TNL (TIR-NBS-LRR) [68]. This classification provides critical insights into potential functional specialization and evolutionary relationships. For tandem duplication analysis, adjacent gene pairs separated by ≤ 8 genes are typically retrieved from genomes, and their relative orientations (head-to-head, head-to-tail, or tail-to-tail) are determined using tools like BEDTools [19]. Statistical significance is evaluated by χ² tests against random expectations through permutation tests (e.g., 10,000 permutations) to distinguish biologically meaningful clusters from random arrangements [19].

Figure 1: Experimental workflow for identifying and analyzing tandem duplication clusters in NBS domain genes

Evolutionary and Expression Analyses

Understanding the evolutionary forces shaping tandem duplication clusters requires integrated analyses of sequence divergence, selection pressures, and expression patterns. The Ka/Ks calculation serves as a fundamental metric for detecting selection pressures, where Ka/Ks > 1 indicates positive selection, Ka/Ks < 1 signifies purifying selection, and Ka/Ks ≈ 1 suggests neutral evolution [67]. These calculations are typically performed using tools like KaKs_Calculator 3.0 with appropriate transition/transversion ratios [69].

For expression divergence analysis, RNA-seq data from various tissues and stress conditions are processed through transcriptomic pipelines, with expression values (typically FPKM) categorized into tissue-specific, abiotic stress-specific, and biotic-stress-specific profiles [3]. Studies on tandemly duplicated genes in potato have revealed significant correlations among expression, promoter, and protein divergences, providing insights into the mechanisms of functional retention [67]. Orthologous gene analysis using tools like OrthoFinder further enables the identification of conserved NLR gene pairs between species, revealing genes preserved during domestication processes and lineage-specific expansions [19].

Experimental Data: Insights from Key Plant Systems

Tandem Duplication Patterns in Crop Genomes

Comprehensive genome-wide analyses across multiple crop species have revealed the profound impact of tandem duplication on the evolution of NBS domain genes. In potato genotypes, tandemly duplicated genes are abundant and dispersed throughout the genome, with several functional specificities differentially enriched across genomes, including disease resistance, stress tolerance, and biosynthetic pathways [67]. Approximately one-fourth of tandemly duplicated gene clusters are lineage-specific among multiple potato genomes, and these tend to localize toward centromeres while revealing distinct selection signatures and expression patterns [67].

The pepper genome demonstrates uneven distribution of NBS-LRR genes across chromosomes, with chromosome 3 harboring the highest number of genes (38) and containing the largest gene cluster comprising eight genes [66]. This distribution highlights the uneven localization and clustering patterns of resistance genes, with chromosome 12 exhibiting the highest diversity of gene subclasses while chromosomes 2 and 6 contain the lowest gene numbers [66]. These clustering patterns reflect the dynamic interplay between tandem duplication and genomic rearrangements in shaping resistance gene evolution.

Table 2: Tandem Duplication Characteristics in Plant Genomes

Plant Species	Lineage-Specific Clusters	Genomic Distribution	Functional Enrichment	Evolutionary Fate
Potato Genotypes	~25% of TDG clusters	Dispersed, centromeric bias	Disease resistance, Stress tolerance, Biosynthetic pathways	Sub-functionalization dominant [67]
Pepper	47 clusters across genome	Chromosome 3: highest density	Pathogen recognition, Immune signaling	nTNL subfamily expansion [66]
Fabaceae Crops	Species-specific clusters	Variable across species	Preferential NB-ARC & LRR co-occurrence	Species-specific diversification [68]
Garden Asparagus	16 conserved NLR pairs with wild relative	Chromosomal clustering	Defense responses, Phytohormone signaling	Contraction during domestication [19]

Sequence Divergence and Functional Specialization

Sequence divergence among tandemly duplicated NBS genes follows distinct patterns that reflect their functional specialization. Structural analyses of pepper NBS-LRR genes have identified six conserved motifs (P-loop, RNBS-A, kinase-2, RNBS-B, RNBS-C, and GLPL) essential for ATP/GTP binding and resistance signaling [66]. Subfamily-specific differences in motif composition and sequence similarity highlight their functional divergence and specialization, with the central NB-ARC domain containing conserved motifs critical for immune function [19].

Studies in potato genotypes have revealed that the majority of duplicated genes are retained through sub-functionalization followed by genetic redundancy, while only a small fraction of duplicated genes is retained through neo-functionalization [67]. This pattern indicates that conservation of existing functions rather than acquisition of novel functions represents the primary evolutionary trajectory for most tandemly duplicated NBS genes. The expression divergence between duplicated genes is significantly correlated with both promoter and protein sequence divergence, suggesting coordinated evolution of regulatory and coding sequences in shaping functional diversity [67].

Table 3: Essential Research Reagents and Computational Tools for NBS Gene Analysis

Tool/Resource	Type	Primary Function	Application Context
OrthoFinder	Computational Tool	Orthogroup inference and phylogenetic analysis	Evolutionary relationships of NBS genes across species [3]
tRNAscan-SE	Computational Tool	tRNA gene identification	Genome annotation and duplication analysis [69]
PRGminer	Deep Learning Tool	R-gene prediction and classification	High-throughput identification of resistance genes [38]
InterProScan	Domain Database	Protein domain identification	Classification of NBS gene architectures [46] [19]
KaKs_Calculator	Evolutionary Tool	Ka/Ks calculation	Selection pressure analysis on duplicated genes [69] [67]
MEME Suite	Motif Analysis	Conserved motif discovery	Identification of NBS domain motifs [19] [66]
PlantCARE	Database	cis-acting element analysis	Promoter analysis of NBS genes [19]
Phytozome	Genomic Database	Plant genome data repository	Source of genomic sequences and annotations [3] [69]

The comparative analysis of tandem duplication clusters and sequence divergence in NBS domain genes reveals fundamental principles governing plant immunity evolution. The persistent pattern of lineage-specific expansion through tandem duplication across diverse species underscores its crucial role in generating functional diversity for pathogen recognition [3] [67]. This evolutionary mechanism provides plant genomes with adaptable genetic toolkits to counter rapidly evolving pathogens, with sub-functionalization serving as the dominant pathway for retaining duplicated genes while maintaining genomic stability [67].

For crop improvement programs, these insights offer strategic guidance for disease resistance breeding. The identification of conserved NLR gene pairs between wild and cultivated species, as demonstrated in asparagus, highlights potential candidates for introgression breeding [19]. Similarly, the recognition that tandemly duplicated genes are frequently enriched in disease resistance functions suggests that genomic regions with high cluster density represent priority targets for marker development and gene pyramiding [66] [67]. Emerging deep learning tools like PRGminer, which achieves 95-98% accuracy in R-gene prediction, further enhance our capacity to identify valuable resistance genes across diverse germplasm [38].

As genomic technologies continue to advance, the integration of comparative genomic analyses with functional validation through methods like virus-induced gene silencing (VIGS) will be essential for translating insights from tandem duplication studies into practical crop improvement strategies [3]. This integrated approach promises to accelerate the development of durable disease resistance in agricultural systems facing evolving pathogen threats.

Resolving Complex Phylogenies and Ortholog Delineation

In the study of plant disease resistance, the NBS-LRR gene family represents one of the largest and most critical classes of resistance (R) genes, forming a fundamental component of the plant immune system [70] [32]. These genes encode intracellular receptors that recognize pathogen-secreted effectors and initiate robust immune responses, often culminating in hypersensitive response and programmed cell death to restrict pathogen spread [46] [32]. The comparative analysis of NBS domain genes across plant species presents substantial computational challenges due to their remarkable diversity, complex genomic architecture, and rapid evolution driven by pathogen pressures [3] [19].

Complex phylogenetic relationships within the NBS-LRR family arise from several biological factors. These genes exhibit extraordinary variation in copy number across plant species, ranging from just 25 NLRs in the bryophyte Physcomitrella patens to over 2,000 in hexaploid wheat (Triticum aestivum) [3] [19]. This expansion occurs primarily through duplication events, with whole-genome duplication (WGD), tandem, and dispersed duplications all contributing to the rapid evolution of this gene family [70] [3]. Additionally, NBS-LRR genes display diverse domain architectures beyond the typical TIR-NBS-LRR (TNL), CC-NBS-LRR (CNL), and RPW8-NBS-LRR (RNL) configurations, including numerous truncated forms that lack specific domains yet retain functional importance [71] [46].

Ortholog delineation within this complex family is particularly challenging due to several factors. NBS-LRR genes are frequently organized in clusters of closely duplicated genes within plant genomes, complicating accurate annotation and differentiation between recent paralogs and true orthologs [19] [38]. Their characteristically low expression levels and frequent misannotation as repetitive elements further exacerbate these challenges [38]. This comparative guide evaluates the computational strategies and tools available for resolving these complexities, providing researchers with a framework for accurate phylogenetic reconstruction and ortholog identification in NBS domain gene research.

Computational Tools and Methodologies

Tool Classification and Functional Comparison

A diverse array of computational tools has been developed to address the challenges of NBS-LRR gene identification, phylogenetic analysis, and ortholog delineation. These tools employ different methodological approaches, from traditional domain-based searches to modern machine learning frameworks, each with distinct strengths and applications as shown in Table 1.

Table 1: Computational Tools for NBS-LRR Gene Analysis

Tool Name	Primary Function	Methodology	Input Data	Key Features
PRGminer [38]	R-gene prediction & classification	Deep learning	Protein sequences	Dipeptide composition-based; 95.72% accuracy on independent test
OrthoFinder [3] [19]	Orthogroup inference	Graph-based clustering	Protein sequences from multiple species	Uses DIAMOND for sequence similarity, MCL for clustering
MCScanX [70]	Synteny analysis	Comparative genomics	Genomic coordinates & BLAST results	Detects collinear blocks, differentiates duplication types
NLR-Annotator [32]	NLR-specific annotation	Domain-based search	Genomic or protein sequences	Specialized for NBS-LRR gene family
DRAGO2/3 [32]	R-gene prediction	Domain architecture	Protein sequences	Pipeline-based approach
RGAugury [32]	R-gene annotation	Integrated pipeline	Genomic data	Combines multiple prediction methods
KaKs_Calculator [70]	Selection pressure analysis	Evolutionary metrics	Coding sequences	Calculates Ka/Ks ratios using NG model
MEME Suite [71] [6]	Motif discovery	Expectation maximization	Protein sequences	Identifies conserved motifs, width 6-50 aa

The selection of an appropriate tool depends heavily on the research objectives. For comprehensive genome-wide identification of NBS-LRR genes, domain-based approaches using HMMER with the NB-ARC domain (PF00931) remain the gold standard [71] [70]. These methods leverage the conserved nucleotide-binding site that defines this gene family, providing high sensitivity for initial identification. For ortholog delineation across multiple species, OrthoFinder implements a robust graph-based approach that clusters proteins into orthogroups based on sequence similarity, effectively grouping genes that descended from a single gene in the last common ancestor [3]. For evolutionary analysis, MCScanX enables the detection of syntenic blocks across genomes, allowing researchers to distinguish between different types of gene duplications and their contributions to NBS-LRR family expansion [70].

Recent advances in machine learning approaches have shown promising results, particularly for challenging annotation scenarios. PRGminer utilizes deep learning to predict resistance genes based on dipeptide composition, achieving 95.72% accuracy on independent testing data [38]. This method offers advantages when analyzing genomes with low homology to well-characterized reference species, where traditional similarity-based methods may fail. The integration of these complementary approaches provides a powerful toolkit for resolving the complex phylogenies of NBS domain genes.

Experimental Protocols for Phylogenetic and Ortholog Analysis

Genome-Wide Identification Protocol

The initial step in NBS-LRR gene analysis involves comprehensive identification across target genomes. The standard protocol utilizes Hidden Markov Model (HMM) searches with the NB-ARC domain (PF00931) as the query [71] [70]. This process begins with HMMER software (v3.1b2) using a stringent E-value cutoff (typically 1*10⁻²⁰) to identify candidate NBS-containing genes [71]. Following initial identification, candidate sequences undergo domain validation using multiple databases including Pfam, SMART, and NCBI's Conserved Domain Database (CDD) to confirm the presence of characteristic NBS-LRR domains [71] [6]. Additional domains (TIR, CC, LRR, RPW8) are identified using Pfam domains (PF01582, PF00560, PF07723, PF07725, PF12779, PF13306, PF13516, PF13855, PF14580, PF03382, PF01030, PF05725) and coiled-coil domains are confirmed via NCBI CDD [70]. This multi-step verification ensures both sensitivity and specificity in gene identification.

Phylogenetic Reconstruction Protocol

For phylogenetic analysis of identified NBS-LRR genes, the standard workflow involves multiple sequence alignment using either Clustal W or MUSCLE v3.8.31 with default parameters [71] [70]. The aligned sequences then undergo model selection based on the best-fit evolutionary model, often determined by statistical criteria. The tree construction typically employs the maximum likelihood method implemented in MEGA11 or MEGA7, based on models such as Whelan and Goldman + freq. Model or JTT matrix-based model [71] [19]. Statistical support for tree nodes is assessed through bootstrap analysis with 1000 replicates [71]. The resulting phylogenetic trees enable classification of NBS-LRR genes into major clades (TNL, CNL, RNL) and facilitate evolutionary inferences. For example, in Nicotiana benthamiana, this approach classified 156 NBS-LRR homologs into three major clades containing 5 TNL-type, 25 CNL-type, 23 NL-type, 2 TN-type, 41 CN-type, and 60 N-type proteins [71].

Ortholog Delineation Protocol

Ortholog identification across multiple species follows a systematic approach beginning with protein sequence clustering using OrthoFinder v2.5.1, which employs DIAMOND for fast sequence similarity searches and the MCL clustering algorithm for orthogroup assignment [3]. The resulting orthogroups represent sets of genes descended from a single gene in the last common ancestor of the species being compared. To identify evolutionary patterns, syntenic analysis is performed using MCScanX with protein sequences from compared species through reciprocal BLASTP searches [70]. For genes within syntenic blocks, selection pressure is quantified by calculating non-synonymous (Ka) and synonymous (Ks) substitution rates using KaKs_Calculator 2.0 with appropriate evolutionary models such as Nei-Gojobori (NG) [70]. This integrated protocol enables researchers to distinguish true orthologs from recent paralogs and identify genes under positive selection, which may indicate adaptive evolution in response to pathogen pressures.

Table 2: Conserved Motifs in NBS Domain Genes

Motif Name	Conservation Level	Functional Role	Detection Method
P-loop	High	Nucleotide binding	MEME, HMMER
GLPL	High	Domain organization	MEME, HMMER
MHD	High	Regulatory function	MEME, HMMER
Kinase 2	High	Signal transduction	MEME, HMMER
RNBS-A	Medium	Structural motif	MEME
RNBS-B	Medium	Structural motif	MEME
RNBS-C	Medium	Structural motif	MEME
RNBS-D	Medium	Structural motif	MEME

Visualization of Methodological Workflows

Phylogenetic Analysis Workflow

The following diagram illustrates the standardized workflow for phylogenetic analysis of NBS-LRR genes, integrating multiple computational tools and validation steps:

Ortholog Delineation Workflow

The ortholog delineation process involves comparative genomics across multiple species, as visualized in the following workflow:

Performance Comparison and Benchmarking

Tool Performance Metrics

Different computational approaches exhibit varying performance characteristics in NBS-LRR gene analysis. Domain-based methods using HMMER with the NB-ARC domain provide the foundation for most NBS-LRR identification pipelines, offering robust performance for initial gene discovery [71] [70]. These methods successfully identified 156 NBS-LRR homologs in Nicotiana benthamiana representing approximately 0.25% of all annotated genes in the genome [71], and 1226 NBS genes across three Nicotiana genomes (N. tabacum, N. sylvestris, and N. tomentosiformis) [70].

Machine learning approaches demonstrate superior performance in certain scenarios, particularly for challenging annotations. PRGminer achieved 98.75% accuracy in k-fold training/testing and 95.72% on independent testing for Phase I (R-gene vs. non-R-gene classification), with Matthews correlation coefficients of 0.98 and 0.91 respectively [38]. For Phase II (R-gene classification into eight categories), it maintained 97.55% accuracy in k-fold testing and 97.21% on independent testing [38]. These results indicate that deep learning methods can effectively complement traditional approaches, especially for genomes with limited reference annotations.

Ortholog clustering tools like OrthoFinder have been successfully applied to large-scale comparative analyses. One study identified 12,820 NBS-domain-containing genes across 34 species from mosses to monocots and dicots, classifying them into 168 distinct classes with several novel domain architecture patterns [3]. This analysis revealed 603 orthogroups, including both core (widely conserved) and unique (species-specific) orthogroups, with tandem duplications playing a significant role in NBS gene expansion [3].

Case Studies in Plant Lineages

The performance of phylogenetic and ortholog analysis methods varies across plant lineages with different genomic characteristics. In medicinal plants like Salvia miltiorrhiza, researchers identified 196 NBS-LRR genes, among which only 62 possessed complete N-terminal and LRR domains [46]. Comparative analysis revealed a marked reduction in TNL and RNL subfamily members within Salvia species compared to model plants, demonstrating how lineage-specific evolutionary patterns can be uncovered through these methods [46].

In horticultural crops like garden asparagus (Asparagus officinalis) and its wild relatives, comparative genomic analysis revealed a dramatic contraction of NLR genes during domestication, with gene counts of 63, 47, and 27 NLR genes identified in A. setaceus, A. kiusianus, and A. officinalis respectively [19]. Orthologous gene analysis identified only 16 conserved NLR gene pairs between A. setaceus and A. officinalis, representing the NLR genes preserved during domestication [19]. This case study demonstrates how ortholog delineation can reveal important evolutionary trends with practical implications for crop breeding.

Table 3: NBS-LRR Gene Distribution Across Select Plant Species

Plant Species	Total NBS Genes	TNL	CNL	RNL	Atypical	Study/Reference
Nicotiana benthamiana	156	5	25	-	126	[71]
Nicotiana tabacum	603	64	74	-	465	[70]
Akebia trifoliata	73	19	50	4	-	[6]
Salvia miltiorrhiza	196	2	61	1	132	[46]
Asparagus officinalis	27	Not specified	Not specified	Not specified	Not specified	[19]
Triticum aestivum	~2,000	Not specified	Not specified	Not specified	Not specified	[3]

Research Reagent Solutions

Successful phylogenetic and ortholog analysis of NBS domain genes requires specific research reagents and computational resources. The following essential materials represent the core toolkit for researchers in this field:

• HMMER Suite: Software package for sequence homology searches using profile hidden Markov models; essential for identifying NBS-ARC domains (PF00931) with high sensitivity [71] [70].
• OrthoFinder: Phylogenetic orthology inference tool that clusters proteins into orthogroups across multiple species; implements DIAMOND for fast sequence similarity and MCL for clustering [3] [19].
• MEME Suite: Toolset for motif discovery and analysis; used to identify conserved motifs (P-loop, GLPL, MHD, Kinase 2) within NBS domains with motif widths typically set between 6-50 amino acids [71] [6].
• MCScanX: Synteny analysis tool that detects collinear blocks across genomes; differentiates between whole-genome, tandem, and segmental duplications in NBS gene family expansion [70].
• MEGA Software: Molecular Evolutionary Genetics Analysis tool for multiple sequence alignment, model selection, and phylogenetic tree construction using maximum likelihood methods [71] [19].
• KaKs_Calculator: Software for calculating non-synonymous (Ka) and synonymous (Ks) substitution rates; quantifies selection pressure on NBS genes using models like Nei-Gojobori (NG) [70].
• PRGminer: Deep learning-based tool for resistance gene prediction; uses dipeptide composition for classification with 95.72% accuracy on independent testing [38].
• PlantCARE Database: Repository for plant cis-acting regulatory elements; analyzes promoter regions (typically 1500-2000 bp upstream) of NBS-LRR genes for defense-related elements [71] [19].
• Pfam Database: Curated collection of protein families and domains; provides essential HMM profiles for domain identification and verification [71] [70].
• NCBI CDD: Conserved Domain Database for annotation of functional domains in protein sequences; particularly important for identifying coiled-coil domains not always detected by Pfam [70] [6].

The resolution of complex phylogenies and accurate delineation of orthologs in NBS domain gene research requires integrated computational approaches combining established domain-based methods with emerging machine learning techniques. The comparative analysis presented in this guide demonstrates that while HMM-based searches using the NB-ARC domain remain fundamental for initial gene identification, deep learning tools like PRGminer offer complementary advantages for challenging annotation scenarios, particularly in non-model plant species with limited reference genomes [71] [38].

For phylogenetic reconstruction, maximum likelihood methods implemented in MEGA software provide robust frameworks for classifying NBS-LRR genes into evolutionary clades, supported by bootstrap validation [71] [19]. For ortholog delineation across multiple species, OrthoFinder delivers reliable orthogroup clustering, while MCScanX enables the detection of syntenic relationships that reveal patterns of gene family expansion through different duplication mechanisms [70] [3].

The performance benchmarks and case studies presented reveal that the choice of computational strategy should be guided by specific research objectives, considering factors such as taxonomic scope, genomic complexity, and available computational resources. As medicinal plant genomics advances, with over 400 genomes from 203 plants sequenced as of February 2025 [72], these computational approaches will play an increasingly vital role in uncovering the evolutionary dynamics of disease resistance genes and facilitating their application in crop improvement and drug development.

Managing Data Quality and Class Imbalance in ML Models

In the field of plant genomics, the application of machine learning (ML) to study nucleotide-binding site (NBS) domain genes is revolutionizing our understanding of plant immunity. However, the success of these computational models hinges on overcoming two significant challenges: ensuring high-quality input data and managing the inherent class imbalance in biological datasets. This guide provides a comparative analysis of strategies to address these issues, framed within the context of comparative analysis of NBS domain genes across plant species.

The Critical Role of Data Quality and Balance in Genomic Research

High-quality data is the foundation upon which reliable, accurate, and effective machine learning models are built [73]. In genomic studies of NBS domain genes—one of the largest and most variable plant protein families—researchers often encounter severely class-imbalanced datasets [74]. For instance, a model trained to identify rare resistance genes might be presented with thousands of common genes but only a handful of the crucial disease-resistant variants, causing the model to neglect the minority class and provide misleading accuracy scores [75].

The consequences of poor data quality and imbalance are particularly acute in biological research. Models may become biased, overfitting to noise or the majority class, which leads to poor performance when deployed on real-world, unseen genomic data [73]. This can directly impact the identification of key genes, such as those in orthogroups OG2, OG6, and OG15, which have been shown to be upregulated in plants tolerant to cotton leaf curl disease (CLCuD) [3]. Therefore, implementing robust data quality and rebalancing processes is not merely a technical step but a prerequisite for biologically meaningful discovery.

Comparative Analysis of Data Quality Management Techniques

Data quality management ensures that the data used for model training is accurate, complete, and consistent. The following table compares various techniques and tools applicable to genomic data pipelines.

Table 1: Comparative Analysis of Data Quality Management Techniques

Technique Category	Specific Method/Tool	Key Functionality	Applicability to Genomic Data
Anomaly Detection	Isolation Forest [76] [77], One-Class SVM [77]	Identifies outliers or unusual patterns in data.	Detecting sequencing errors or anomalous gene expression values.
Handling Missing Data	Imputer (Scikit-learn/PySpark) [76], K-NN Imputation [73] [77], MICE [77]	Fills in missing values using statistical methods or predictions.	Imputing missing phenotypic data or gaps in sequence alignments.
Deduplication	MLlib (PySpark) [76], Fuzzy String Matching, NLP [77]	Removes duplicate records based on exact or fuzzy matching.	Identifying and merging duplicate gene entries from multiple databases.
Validation & Standardization	Schema Validation [73], Pattern Recognition [77]	Ensures data conforms to expected formats and business rules.	Validating gene identifier formats or standardizing protein domain names.
Automated Monitoring	AI/ML Platforms (e.g., DataBuck) [78]	Provides real-time data quality checks and continuous monitoring.	Monitoring data streams from high-throughput sequencing platforms.

Comparative Analysis of Class Imbalance Handling Strategies

When the class distribution is skewed, specialized strategies are required to prevent model bias. The table below compares common approaches for handling class imbalance, relevant to scenarios like identifying a small number of disease-resistant NBS genes within a large genome.

Table 2: Comparative Analysis of Class Imbalance Handling Strategies

Strategy	Key Principle	Advantages	Disadvantages/Limitations
Cost-sensitive Learning	"Upweighting" the minority class during loss calculation [74].	Simple to implement; does not alter the original data.	Requires careful tuning of the weight parameter.
Oversampling (e.g., Random Oversampling)	Increasing the number of minority class instances by duplication [75].	Balances the dataset without losing any information.	Can lead to overfitting, especially if duplicates dominate.
Undersampling (e.g., Random Undersampling)	Randomly removing instances from the majority class [75].	Reduces dataset size and training time.	May remove potentially useful data from the majority class.
Synthetic Data Generation (e.g., SMOTE)	Generating synthetic examples for the minority class [75].	Increases diversity of the minority class; reduces overfitting.	May generate noisy samples; less effective for high-dimensional data.
Ensemble Methods (e.g., BalancedBaggingClassifier)	Combining multiple learners trained on balanced subsets of data [75].	Often achieves higher performance and robustness.	Computationally more intensive and complex to implement.

Experimental Protocols for Model Validation

To ensure that the chosen strategies for data quality and class imbalance are effective, rigorous experimental validation is essential. The following protocols outline standard methodologies for benchmarking performance.

Protocol for Benchmarking Data Quality Techniques

Objective: To quantitatively evaluate the impact of different data quality treatments on the performance of a model for NBS gene classification.

• Dataset Preparation: Begin with a raw genomic dataset, such as one containing identified NBS-domain-containing genes from a study across 34 plant species [3].
• Introduce Controlled Anomalies: Artificially introduce specific, measurable data quality issues into a clean subset of the data. This includes:
  • Missing Data: Randomly remove a set percentage (e.g., 5%, 10%) of values from specific features.
  • Duplicates: Introduce duplicate gene records with slight variations in identifiers or annotations.
  • Inconsistencies: Alter formatting in categorical fields (e.g., mixed case in domain names like "NBS-LRR" and "nbs-lrr").
• Apply Quality Treatments: Process the corrupted dataset using different techniques from Table 1.
  • Group 1: Apply anomaly detection followed by removal.
  • Group 2: Apply imputation for missing values and deduplication.
  • Control Group: Use the raw, corrupted data without any treatment.
• Model Training and Evaluation: Train an identical baseline classification model (e.g., Random Forest) on each of the treated datasets and the control. Evaluate models on a pristine, held-out test set using metrics from Section 5. The treatment that results in the model with the highest performance metrics is deemed most effective for that data scenario.

Protocol for Benchmarking Class Imbalance Strategies

Objective: To compare the efficacy of various class imbalance strategies in identifying a minority class of NBS genes (e.g., disease-resistant variants).

• Define Imbalanced Dataset: Utilize a genomic dataset with a known, severe class imbalance. An example could be a set of NLR genes from asparagus species, where a small number of genes are associated with resistance to a pathogen like Phomopsis asparagi, and the majority are not [19].
• Apply Rebalancing Strategies: Process the dataset using the different strategies outlined in Table 2. This creates several training sets:
  • Original imbalanced data (Baseline)
  • Randomly oversampled data
  • Randomly undersampled data
  • SMOTE-processed data
  • Data used with a BalancedBaggingClassifier
• Model Training and Evaluation: Train a classifier on each of the prepared training sets. Unlike the data quality experiment, evaluate all models on the original, imbalanced test set to simulate real-world performance. Compare the results using the metrics detailed in Section 5, with a focus on F1-score and recall for the minority class.

Key Performance Metrics for Model Assessment

When dealing with imbalanced datasets, standard metrics like accuracy can be profoundly misleading [75]. It is crucial to adopt a more nuanced set of evaluation criteria.

• Precision: Measures the accuracy of positive predictions. For a gene classifier, it is the proportion of genes predicted as disease-resistant that are actually resistant. High precision means fewer false positives.
• Recall (Sensitivity): Measures the ability to find all positive instances. It is the proportion of truly disease-resistant genes that were correctly identified by the model. High recall means fewer false negatives.
• F1-Score: The harmonic mean of precision and recall. This single metric balances the trade-off between the two and is a preferred benchmark for imbalanced classification problems [75]. It is calculated as: F1 = 2 * (Precision * Recall) / (Precision + Recall)
• Area Under the Receiver Operating Characteristic Curve (AUC-ROC): Measures the model's ability to distinguish between classes across all classification thresholds. A higher AUC indicates better overall performance.

Workflow Visualization: An Integrated ML Pipeline for NBS Gene Analysis

The following diagram synthesizes the concepts of data quality management and class imbalance handling into a cohesive workflow for NBS gene research.

Successful implementation of the aforementioned strategies requires a suite of computational tools and databases.

Table 3: Essential Research Reagents and Resources for Computational Analysis

Resource Category	Specific Tool / Database	Function in Research
Bioinformatics Pipelines	OrthoFinder [3] [19], HMMER [3] [19], InterProScan [19] [32]	Ortholog clustering, domain identification, and functional annotation of NBS genes.
Genomic Databases	PRGdb [19], ANNA: Angiosperm NLR Atlas [3], Plaza [3]	Provide curated collections of known resistance genes and genomic data for comparative analysis.
Machine Learning Libraries	Scikit-learn [73] [76], Imbalanced-learn (imblearn) [75], PySpark MLlib [76]	Offer implementations of data preprocessing, classification algorithms, and resampling techniques.
Programming Environments	Python, R	Provide the foundational ecosystem for data manipulation, statistical analysis, and model development.

Interpreting the Functional Impact of LRR Domain Loss

The Leucine-Rich Repeat (LRR) domain is a critical structural and functional component of numerous proteins involved in immune recognition and signaling pathways across plant and animal kingdoms. In plants, the nucleotide-binding site (NBS)-LRR gene family constitutes the largest and most prominent class of disease resistance (R) genes, with the LRR domain playing a pivotal role in pathogen recognition and subsequent immune activation [79] [32]. The functional impact of LRR domain loss represents a significant area of research with profound implications for understanding evolutionary biology, host-pathogen interactions, and disease susceptibility. This phenomenon is observed across various species and has been linked to both adaptive evolutionary strategies and increased vulnerability to pathogens.

The structural integrity of NBS-LRR proteins is essential for their function as intracellular immune receptors. These proteins typically consist of three core domains: a variable N-terminal domain (often TIR or CC), a central nucleotide-binding site (NBS) domain, and a C-terminal LRR domain [79]. The LRR domain, characterized by repetitive sequences rich in leucine residues, forms a curved solenoid structure that facilitates protein-protein and protein-ligand interactions, enabling specific recognition of pathogen effectors [80]. This review comprehensively examines the functional consequences of LRR domain loss through comparative genomic analyses, experimental validations, and evolutionary studies across diverse plant species, providing insights into the complex balance between immune system maintenance and evolutionary adaptation.

Structural and Functional Roles of the LRR Domain

Molecular Architecture and Recognition Mechanisms

The LRR domain exhibits a conserved structural architecture that underlies its biological functions. Typically composed of multiple repeats of 20-30 amino acid sequences, each repeat contributes to the formation of a parallel β-sheet on the concave surface and various secondary structures on the convex surface, creating a distinctive horseshoe-like shape [80]. This structural arrangement provides an extensive surface for molecular interactions, with the variable residues in the β-sheet region determining specificity for pathogen recognition. In NLR proteins, the LRR domain commonly consists of multiple repeating units; for instance, in the human NLRP3 inflammasome, the LRR domain comprises 11 repetitive patterns, each containing approximately 28-29 amino acids with a consensus sequence of xLxxLxLxxN/CxLxxxxxxxLxxxLxxxxx [80].

The LRR domain serves multiple critical functions in immune signaling pathways. In plant NBS-LRR proteins, it primarily acts as the molecular sensor that directly or indirectly recognizes pathogen-derived effector proteins, initiating effector-triggered immunity (ETI) [79] [32]. Additionally, the LRR domain participates in autoinhibition and regulation of the protein's activation state. In the resting state, the LRR domain maintains the protein in an inactive conformation, while upon pathogen recognition, conformational changes enable nucleotide exchange in the NBS domain, triggering downstream signaling cascades [80]. This regulatory function highlights the dual role of the LRR domain in both pathogen perception and immune activation control.

LRR Domain in Signaling Complex Assembly

Beyond pathogen recognition, the LRR domain facilitates the assembly of multiprotein signaling complexes. In plant immunity, activated NBS-LRR proteins often form resistosomes or signaling hubs that initiate downstream defense responses, including hypersensitive response (HR) and programmed cell death [32]. Similarly, in mammalian systems, the LRR domain of NLRP3 contributes to inflammasome assembly by facilitating oligomerization and recruitment of adapter proteins such as ASC and caspase-1 [80]. The structural versatility of the LRR domain enables its participation in diverse protein-protein interactions, allowing immune receptors to integrate signals from multiple pathways and mount appropriate defense responses tailored to specific pathogen challenges.

Table 1: Key Functions of LRR Domains in Immune Receptors

Function	Mechanism	Biological Significance
Pathogen Recognition	Direct or indirect binding to pathogen effectors through variable residues in β-sheet regions	Specific immunity against diverse pathogens
Signal Regulation	Maintaining autoinhibition in resting state; conformational changes upon activation	Prevention of autoimmunity; controlled immune activation
Complex Assembly	Facilitating oligomerization and recruitment of downstream signaling components	Amplification of immune signals; coordination of defense responses
Subcellular Localization	Interaction with cellular membranes or organelles	Spatial regulation of immune signaling

Genomic and Evolutionary Evidence of LRR Domain Loss

Comparative Genomic Analyses Across Plant Species

Comparative genomic studies have revealed that LRR domain loss is a widespread phenomenon in plant evolution, with significant implications for disease resistance capabilities. A comprehensive analysis of NBS-domain-containing genes across 34 plant species, from mosses to monocots and dicots, identified 12,820 genes with considerable diversity in domain architecture, including numerous instances of truncated proteins lacking complete LRR domains [3]. These findings suggest that domain loss represents an evolutionary strategy for immune system diversification and adaptation.

In the Asparagus genus, a marked contraction of NLR genes was observed during domestication, with wild relatives (A. setaceus: 63 NLRs; A. kiusianus: 47 NLRs) possessing significantly more NLR genes than cultivated garden asparagus (A. officinalis: 27 NLRs) [19]. This reduction in gene number was accompanied by functional impairments, as evidenced by distinct phenotypic responses to pathogen inoculation: A. officinalis was susceptible, while A. setaceus remained asymptomatic. Expression analysis further revealed that most preserved NLR genes in the cultivated species showed either unchanged or downregulated expression following fungal challenge, indicating that domestication selected for reduced investment in disease resistance at both genetic and regulatory levels [19].

The Case of Cultivated Peanut: Balancing Domain Loss and Gene Birth

A particularly instructive example of LRR domain loss comes from studies of cultivated peanut (Arachis hypogaea cv. Tifrunner). Comparative analysis with its diploid ancestors (A. duranensis and A. ipaensis) revealed that although the cultivated tetraploid possesses more full-length NBS-LRR genes (713) than its progenitors (A. duranensis: 278; A. ipaensis: 303), these genes contain fewer LRR domains [28]. This reduction in LRR domain content was associated with relaxed selection pressure on LRR domains and NBS-LRR proteins in the cultivated species.

Quantitative trait locus (QTL) analysis connected this domain loss to the crop's disease resistance profile. Among 113 NBS-LRRs associated with disease resistance QTLs in cultivated peanut, 75 were classified as "young" genes (originating after tetraploidization) while only 38 were "old" genes (inherited from progenitors) [28]. This finding suggests that recent gene birth partially compensates for LRR domain loss. However, the overall reduction in LRR domain content provides a compelling explanation for the greater susceptibility of cultivated peanut to diseases compared to its wild relatives, illustrating the complex evolutionary trade-offs between genome streamlining and maintaining immune competence.

Table 2: Examples of LRR Domain Loss in Different Plant Species

Species	NLR Gene Count	LRR Domain Status	Functional Consequences
Arachis hypogaea cv. Tifrunner (cultivated peanut)	713 full-length NBS-LRRs	Fewer LRR domains compared to diploid progenitors	Reduced disease resistance despite higher gene number
Asparagus officinalis (garden asparagus)	27 NLR genes	Contraction of NLR repertoire including LRR-containing genes	Increased susceptibility to fungal diseases
Salvia miltiorrhiza (danshen)	196 NBS-containing genes, only 62 with complete LRR domains	High proportion of truncated forms lacking LRR domains	Unknown, but suggests adaptation to specific pathogen pressures

Methodologies for Studying LRR Domain Loss

Genomic Identification and Annotation of NLR Genes

The comprehensive identification of NBS-LRR genes and characterization of their domain architecture form the foundation for studying LRR domain loss. The standard methodological approach involves a multi-step bioinformatics pipeline beginning with genome-wide scans for genes containing the NB-ARC domain (Pfam: PF00931) using Hidden Markov Model (HMM)-based searches with tools like HMMER [3] [19] [57]. Candidate sequences identified through this initial screen are subsequently validated through domain architecture analysis using InterProScan and NCBI's Batch CD-Search to confirm the presence of characteristic NBS and LRR domains [19] [57].

Following identification, NBS-LRR genes are classified based on their domain composition. Typical NLRs contain all three core domains (N-terminal, NBS, and LRR), while atypical forms include various truncated versions such as NL (NBS-LRR, lacking a complete N-terminal domain), N (NBS only), TN (TIR-NBS), and CN (CC-NBS) [57]. This classification enables researchers to quantify the prevalence of LRR domain loss and investigate its functional implications through comparative analysis across species or genotypes with differing disease resistance profiles.

Evolutionary and Expression Analysis

To understand the evolutionary dynamics driving LRR domain loss, researchers employ phylogenetic reconstruction and orthologous gene clustering. Phylogenetic trees are typically constructed using maximum likelihood methods based on aligned NBS domain sequences, allowing visualization of evolutionary relationships and identification of lineage-specific domain loss events [81] [19]. OrthoFinder is commonly used to cluster NLR genes into orthogroups, facilitating the identification of conserved versus lineage-specific genes and the inference of gene duplication and loss events [3].

Expression profiling through RNA-seq analysis under various conditions, including pathogen challenge, provides insights into the functional consequences of LRR domain loss. Studies typically compare expression patterns of intact versus truncated NLR genes in resistant and susceptible genotypes, often revealing differential regulation associated with domain composition [3] [19]. For instance, in asparagus, most preserved NLR genes in the susceptible cultivated species showed either unchanged or downregulated expression following fungal challenge, suggesting that LRR domain loss may be accompanied by regulatory changes that further compromise immunity [19].

Diagram 1: Experimental workflow for studying LRR domain loss, integrating bioinformatic and functional approaches.

Functional Consequences of LRR Domain Loss

Impact on Disease Resistance Profiles

The loss of LRR domains in NLR proteins has demonstrable effects on plant disease resistance, as evidenced by multiple comparative studies. In cultivated peanut, the reduction in LRR domain content correlates with heightened susceptibility to various pathogens, including those causing late leaf spot, tomato spotted wilt virus, and bacterial wilt [28]. QTL mapping identified 113 NBS-LRRs associated with disease resistance in cultivated peanut, with the majority (75 genes) representing young genes that emerged after tetraploidization. This suggests that while new genes can partially compensate for LRR domain loss, the overall reduction in LRR diversity constrains the plant's capacity to recognize and respond to diverse pathogen effectors.

The functional impact of LRR domain loss extends beyond simple quantitative reductions in resistance to include qualitative changes in defense responses. In asparagus, the contraction of the NLR gene repertoire during domestication, which involved both gene loss and LRR domain truncation, resulted not only in increased susceptibility but also in altered expression patterns of retained NLR genes following pathogen challenge [19]. Unlike their wild relatives, cultivated asparagus accessions showed inconsistent induction of NLR genes in response to infection, indicating that domain loss may disrupt regulatory networks coordinating immune responses.

Molecular and Mechanistic Implications

At the molecular level, LRR domain loss impairs critical protein functions essential for effective immunity. The LRR domain facilitates specific recognition of pathogen effectors through direct or indirect binding, with its solvent-exposed residues undergoing diversifying selection to generate recognition specificity [79]. Loss of these domains eliminates crucial binding surfaces, diminishing the plant's capacity to detect invading pathogens. Additionally, LRR domains contribute to proper protein folding, oligomerization, and subcellular localization—all essential for NLR function [80]. Truncated proteins lacking LRR domains may fail to assemble into functional signaling complexes, compromising downstream defense activation.

LRR domain loss also affects signal transduction mechanisms within immune pathways. In intact NLR proteins, the LRR domain interacts with the NBS domain to maintain autoinhibition in the absence of pathogens [80]. Pathogen perception induces conformational changes that relieve this inhibition, enabling nucleotide exchange and activation of downstream signaling. Truncated proteins lacking LRR domains may exhibit altered regulation, potentially resulting in either constitutive activation (leading to autoimmunity) or failure to activate appropriately upon infection. This delicate balance explains why LRR domain loss is often associated with either enhanced susceptibility or, in rare cases, with autoactive variants that trigger defense responses in the absence of pathogens.

Experimental Validation of LRR Domain Function

Functional Characterization Through Genetic Approaches

Several experimental approaches have been employed to validate the functional significance of LRR domains and characterize the consequences of their loss. Virus-induced gene silencing (VIGS) has proven particularly valuable for functional analysis, as demonstrated in cotton studies where silencing of specific NBS genes (e.g., GaNBS from orthogroup OG2) compromised resistance to cotton leaf curl disease [3]. This approach allows rapid assessment of gene function without the need for stable transformation, facilitating high-throughput functional screening of NLR genes.

Protein interaction studies provide direct evidence for the role of LRR domains in pathogen recognition and signal complex formation. In cotton, protein-ligand and protein-protein interaction assays revealed strong binding between specific NBS proteins and both ADP/ATP and core proteins of the cotton leaf curl disease virus [3]. Such interactions typically depend on intact LRR domains, and their disruption through domain loss or mutation abrogates binding capacity. These molecular analyses complement genetic studies by elucidating the mechanistic basis for impaired immunity resulting from LRR domain loss.

Expression Analysis and Regulatory Consequences

Comprehensive expression profiling under various conditions represents another key approach for validating the functional impact of LRR domain loss. Studies in multiple species have analyzed NLR gene expression across different tissues and in response to diverse biotic and abiotic stresses, revealing distinct patterns between intact and truncated NLR genes [3] [19]. For instance, in asparagus, comparative transcriptomic analysis of wild and cultivated species following pathogen inoculation showed that preserved NLR genes in the susceptible cultivated accession exhibited blunted induction compared to their wild counterparts [19].

Promoter analysis of NLR genes has identified numerous cis-elements responsive to defense signals and phytohormones, suggesting complex regulatory networks that may be disrupted by domain loss [19] [57]. Studies in Salvia miltiorrhiza revealed an abundance of cis-acting elements in NBS gene promoters related to plant hormones and abiotic stress, indicating integrated regulation of defense responses that may be compromised in truncated genes [57]. These findings suggest that LRR domain loss may not only affect protein function directly but also indirectly influence gene regulation and expression dynamics.

Table 3: Key Research Reagents and Tools for Studying LRR Domain Loss

Research Tool	Application	Function in Research
HMMER (with Pfam NB-ARC domain PF00931)	Genomic identification	Identification of NBS-containing genes in genome sequences
InterProScan / NCBI CD-Search	Domain annotation	Verification and classification of protein domains
OrthoFinder	Evolutionary analysis	Clustering of NLR genes into orthogroups
MEME Suite	Motif analysis	Identification of conserved motifs in NBS and LRR domains
Virus-Induced Gene Silencing (VIGS)	Functional validation	Transient knockdown of target NLR genes
RNA-seq / Expression profiling	Transcriptomic analysis	Assessment of gene expression under various conditions

The loss of LRR domains in NBS-LRR genes represents a significant evolutionary phenomenon with profound implications for plant immunity and disease resistance. Comparative genomic analyses across diverse plant species have revealed that LRR domain loss is widespread, occurring through both gene contraction and the proliferation of truncated forms. The functional consequences of this domain loss are complex and context-dependent, ranging from compromised disease resistance in cultivated species to potential adaptive advantages in specific environments.

The evidence from multiple systems indicates that LRR domain loss frequently correlates with reduced disease resistance, as demonstrated in cultivated peanut and asparagus, where domain reduction parallels increased susceptibility to pathogens. However, this loss may be partially compensated by the birth of new genes and functional diversification of retained NLRs. The methodological approaches for studying LRR domain loss—spanning genomic identification, phylogenetic analysis, expression profiling, and functional validation—provide powerful tools for elucidating the evolutionary drivers and functional consequences of this phenomenon.

Understanding the impact of LRR domain loss extends beyond academic interest to practical applications in crop improvement. By identifying the specific domains and residues critical for pathogen recognition and immune activation, researchers can develop more precise strategies for enhancing disease resistance through molecular breeding or genetic engineering. Furthermore, recognizing the evolutionary trade-offs between genome streamlining and immune competence informs conservation efforts for wild relatives that serve as reservoirs of genetic diversity for crop improvement programs.

Functional Validation and Comparative Genomics of Disease Resistance

Virus-Induced Gene Silencing (VIGS) for Functional Characterization

Virus-Induced Gene Silencing (VIGS) has emerged as a powerful reverse genetics tool for rapidly characterizing gene function in plants. This technology leverages the plant's innate post-transcriptional gene silencing (PTGS) machinery, using recombinant viral vectors to trigger systemic suppression of endogenous gene expression. The resulting phenotypic changes enable researchers to link gene sequences to biological functions without the need for stable transformation, which is particularly valuable for species with long life cycles or recalcitrant genetic systems [82]. In the specific context of plant immunity, VIGS has become indispensable for functional analysis of nucleotide-binding site (NBS) domain genes, which constitute the largest class of plant resistance (R) proteins. These NBS-leucine rich repeat (NLR) genes play critical roles in effector-triggered immunity by recognizing pathogen-secreted effectors and activating robust defense responses [57]. The integration of VIGS with comparative genomics of NLR genes across species has significantly accelerated the pace of discovery in plant disease resistance mechanisms.

The biological foundation of VIGS lies in the plant's antiviral defense system. When a recombinant viral vector containing a fragment of a host gene is introduced, the plant processes the viral double-stranded RNA replication intermediates into 21-24 nucleotide small interfering RNAs (siRNAs) using Dicer-like enzymes. These siRNAs are incorporated into the RNA-induced silencing complex (RISC), which guides sequence-specific degradation of complementary endogenous mRNA transcripts, thereby knocking down target gene expression [82]. This mechanism enables researchers to study gene function by observing the phenotypic consequences of gene silencing, typically within 2-4 weeks after inoculation, making VIGS significantly faster than traditional stable transformation approaches.

Comparative Analysis of VIGS Systems for NBS Gene Characterization

Key Viral Vector Systems and Their Applications

Multiple viral vectors have been developed for VIGS applications, each with distinct advantages and limitations for functional genomics research. The selection of an appropriate vector system is critical for successful gene silencing, particularly when working with NBS domain genes that often exhibit complex expression patterns and functional redundancy.

Table 1: Major Viral Vectors Used in VIGS for Plant Immunity Research

Vector Type	Virus Origin	Host Range	Silencing Duration	Key Advantages	Primary Limitations	Application in NBS Gene Studies
TRV	Tobacco Rattle Virus	Broad (Solanaceae, Arabidopsis, etc.)	3-8 weeks	Mild symptoms, efficient meristem silencing [82]	Bipartite genome requires two constructs	Functional validation of NBS genes in tomato, tobacco, pepper [82]
BSMV	Barley Stripe Mosaic Virus	Monocots (barley, wheat)	2-6 weeks	Effective in cereal crops [83]	Can cause noticeable symptoms	Characterization of leaf stripe resistance genes in barley [83]
BPMV	Bean Pod Mottle Virus	Soybean	4-8 weeks	High efficiency in soybean [84]	Requires particle bombardment	Analysis of soybean cyst nematode resistance [84]
CLCrV	Cotton Leaf Crumple Virus	Cotton	3-6 weeks	Optimized for cotton species	Limited to malvaceae family	Validation of GaNBS in cotton leaf curl disease resistance [3]

Performance Comparison Across Plant Systems

The effectiveness of VIGS varies significantly across plant species due to differences in viral susceptibility, systemic movement, and RNAi machinery efficiency. For NBS gene characterization, successful applications have been demonstrated in multiple plant families:

In dicotyledonous plants, TRV-based systems have shown particularly broad utility. Research in cotton demonstrated that silencing of GaNBS (orthogroup OG2) through VIGS increased plant susceptibility to cotton leaf curl disease, confirming its functional role in virus resistance [3]. In pepper (Capsicum annuum L.), TRV-VIGS has been successfully employed to characterize genes controlling disease resistance and unique metabolic pathways, providing crucial functional data for this genetically recalcitrant species [82].

For monocotyledonous species, BSMV-based vectors have proven most effective. In barley, BSMV-VIGS was used to validate the function of HvLRR8-1, a leucine-rich repeat receptor-like kinase gene containing an STKc domain, in resistance to Pyrenophora graminea, the causal agent of barley leaf stripe [83]. The system achieved sufficient silencing efficiency to produce measurable changes in disease susceptibility, enabling reliable functional annotation.

Recent optimization work in soybean has addressed previous limitations of VIGS application. An improved TRV-VIGS protocol utilizing Agrobacterium tumefaciens-mediated infection through cotyledon nodes achieved silencing efficiencies of 65% to 95%, successfully knocking down phytoene desaturase (GmPDS), the rust resistance gene GmRpp6907, and the defense-related gene GmRPT4 [84]. This represents a significant advancement for rapid gene validation in legume species.

Experimental Protocols for VIGS in NBS Gene Characterization

Vector Construction and Agroinfiltration Methodology

The fundamental protocol for implementing VIGS involves multiple critical steps that must be optimized for each plant system. The following methodology synthesizes approaches from recent successful applications:

Step 1: Target Gene Fragment Selection and Vector Construction

• Amplify 150-300 bp gene-specific fragment from target NBS gene using cDNA template
• Design primers with appropriate restriction sites (e.g., EcoRI and XhoI for TRV2 vector)
• Clone fragment into viral vector (e.g., pTRV2) using restriction digestion and ligation
• Transform recombinant plasmid into Agrobacterium tumefaciens strain (e.g., GV3101) [84]

Table 2: Essential Research Reagents for VIGS Implementation

Reagent/Category	Specific Examples	Function in VIGS Protocol	Considerations for NBS Gene Studies
Viral Vectors	pTRV1, pTRV2, pBSMV, pBPMV	Deliver target gene sequences into plant cells	Select based on host compatibility; TRV for broad dicot application
Agrobacterium Strains	GV3101, LBA4404	Mediate vector delivery into plant tissues	Optimization of optical density (OD600 = 0.3-1.0) critical for efficiency
Selection Antibiotics	Kanamycin, Rifampicin	Maintain plasmid integrity in bacterial cultures	Use appropriate concentrations for vector and strain
Infiltration Buffers	Acetosyringone, MES, MgCl2	Enhance Agrobacterium infection efficiency	10 mM MgCl2 with 150 μM acetosyringone commonly used
Positive Control Constructs	PDS (phytoene desaturase)	Validate silencing system functionality	Photobleaching phenotype confirms successful VIGS
Negative Control Constructs	Empty vector, GFP	Account for nonspecific effects	Essential for proper interpretation of silencing phenotypes

Step 2: Agroinfiltration and Plant Inoculation

• Grow Agrobacterium cultures containing both viral components (e.g., TRV1 and TRV2-target) to OD600 = 0.4-0.8
• Resuspend cells in infiltration buffer (10 mM MgCl2, 10 mM MES, 150 μM acetosyringone)
• Mix cultures containing complementary viral vectors in 1:1 ratio
• For soybean and difficult-to-transform species, use cotyledon node immersion for 20-30 minutes [84]
• For Nicotiana benthamiana and other amenable species, use syringe infiltration or vacuum infiltration
• Maintain inoculated plants at 19-22°C with high humidity for 48-72 hours to facilitate infection

Step 3: Silencing Validation and Phenotypic Assessment

• Monitor positive control (PDS) plants for photobleaching symptoms at 2-3 weeks post-inoculation
• Quantify target gene expression reduction using qRT-PCR with gene-specific primers
• Challenge silenced plants with appropriate pathogens 3-4 weeks post-inoculation
• Assess disease symptoms, pathogen biomass, and defense marker gene expression [83]

Pathway Diagram: VIGS Mechanism and NBS Gene Function Analysis

Diagram Title: VIGS Mechanism for NBS Gene Function Analysis

Applications in Comparative Analysis of NBS Domain Genes

Case Studies in Diverse Plant Species

VIGS has enabled functional comparisons of NBS domain genes across multiple plant species, revealing both conserved and specialized immune functions:

In cotton (Gossypium hirsutum), comparative analysis identified 12,820 NBS-domain-containing genes across 34 plant species. Expression profiling revealed putative upregulation of specific orthogroups (OG2, OG6, OG15) in different tissues under various biotic and abiotic stresses. VIGS-mediated silencing of GaNBS (OG2) in resistant cotton demonstrated its critical role in limiting cotton leaf curl disease virus titer, establishing a direct functional link between this NBS gene and viral resistance [3].

In tung tree (Vernicia species), genome-wide analysis identified 239 NBS-LRR genes across resistant (V. montana) and susceptible (V. fordii) species. The orthologous gene pair Vf11G0978-Vm019719 showed distinct expression patterns, with the V. montana gene exhibiting upregulated expression during Fusarium wilt infection. VIGS experiments confirmed that Vm019719, activated by VmWRKY64, confers resistance to Fusarium wilt, while its allelic counterpart in susceptible V. fordii carries a promoter deletion that renders it ineffective [48].

In asparagus (Asparagus officinalis), comparative genomic analysis revealed significant contraction of the NLR gene repertoire during domestication, with wild relative A. setaceus containing 63 NLR genes compared to only 27 in cultivated A. officinalis. This reduction, coupled with inconsistent induction of retained NLR genes following pathogen challenge, explains the increased disease susceptibility of domesticated asparagus [19].

Experimental Workflow for Cross-Species NBS Gene Analysis

Diagram Title: Cross-Species NBS Gene Analysis Workflow

Technical Considerations and Optimization Strategies

Critical Factors Influencing VIGS Efficiency

Successful implementation of VIGS for NBS gene characterization requires careful optimization of several parameters:

Insert Design and Vector Selection: For NBS genes, which often belong to large gene families with high sequence similarity, specificity of the target fragment is paramount. Fragments of 150-300 bp with moderate GC content (40-60%) typically yield optimal results. Bioinformatics analysis using tools like OrthoFinder should precede experimental work to ensure fragment specificity, particularly for distinguishing between closely related NBS paralogs [3]. The choice of viral vector must align with both the host plant species and the specific tissues being studied, with TRV providing broad applicability across dicot species and BSMV preferred for monocots.

Agroinfiltration Methodology: Efficiency of Agrobacterium-mediated delivery varies significantly across species. While simple syringe infiltration suffices for Nicotiana benthamiana, optimized protocols involving cotyledon node immersion have demonstrated 80-95% infection efficiency in soybean [84]. Critical parameters include Agrobacterium optical density (OD600 = 0.3-1.0), acetosyringone concentration (100-200 μM), and surfactant inclusion (0.01-0.05% Silwet L-77) for challenging species.

Environmental Conditions: Post-inoculation environmental control significantly impacts silencing efficiency and duration. Maintaining temperatures of 19-22°C for 48-72 hours post-inoculation enhances viral spread while minimizing plant stress responses. Extended lower temperature maintenance (21-23°C) throughout the experiment prolongs silencing duration, particularly important for slow-developing disease phenotypes associated with NBS-mediated resistance [82].

Limitations and Alternative Approaches

While VIGS provides unprecedented speed for gene function analysis, several limitations must be considered. Silencing efficiency varies across tissues, with meristematic regions often showing reduced silencing. The transient nature of VIGS may not be suitable for studying late developmental stages, and incomplete silencing can complicate interpretation for essential genes. For NBS genes in particular, functional redundancy within large gene families may mask phenotypic effects when single genes are silenced.

Complementary approaches include stable transformation for constitutive or tissue-specific overexpression, CRISPR/Cas9 for targeted mutagenesis, and heterologous expression systems for detailed biochemical characterization. The integration of VIGS with multi-omics technologies represents a powerful future direction, enabling correlation of phenotypic changes with transcriptomic, proteomic, and metabolomic perturbations following NBS gene silencing.

VIGS has established itself as an indispensable tool for functional characterization of NBS domain genes, enabling rapid validation of candidate resistance genes identified through comparative genomics. The technology's unique advantage lies in its ability to bridge computational predictions and biological function, particularly valuable for species with complex genomes or challenging transformation systems. As viral vectors continue to be optimized and protocols refined for additional species, VIGS will play an increasingly central role in elucidating the evolutionary dynamics and functional specialization of NBS gene families across the plant kingdom. This knowledge provides the foundation for targeted crop improvement through both traditional breeding and modern biotechnological approaches, ultimately contributing to enhanced agricultural sustainability and food security.

Protein-Ligand and Protein-Protein Interaction Studies

The nucleotide-binding site (NBS) domain is a critical component of plant immune receptors, forming the central nucleotide-binding adaptor shared by APAF-1, R proteins, and CED-4 (NB-ARC) in nucleotide-binding leucine-rich repeat (NLR) proteins [3] [57]. These intracellular immune receptors recognize pathogen effectors and initiate effector-triggered immunity (ETI), providing plants with specific resistance to diverse pathogens [85] [86]. The NBS domain functions as a molecular switch, hydrolyzing ATP/GTP to provide energy for immune signaling activation [57] [28]. The C-terminal leucine-rich repeat (LRR) domain facilitates protein-ligand and protein-protein interactions, playing crucial roles in pathogen recognition specificity [5] [28]. The functional characterization of NBS domain genes relies heavily on methodologies that investigate their interaction profiles with various ligands and partner proteins. This guide provides a comparative analysis of experimental and computational approaches for studying NBS domain protein interactions, supporting research in plant immunity and disease resistance breeding.

Methodological Comparison for Protein Interaction Studies

Experimental Approaches for Interaction Validation

Table 1: Experimental Methods for Protein Interaction Studies

Method	Principle	Applications in NBS Research	Key Metrics	References
Virus-Induced Gene Silencing (VIGS)	Gene silencing through viral vectors delivering target sequences	Functional validation of NBS-LRR genes in disease resistance; e.g., Vm019719 in Vernicia montana against Fusarium wilt	Disease susceptibility index, pathogen titer quantification	[5] [3]
Surface Plasmon Resonance (SPR)	Real-time biomolecular interaction analysis via refractive index changes	Measurement of binding kinetics between nanobodies and ligands	Affinity constants (KD), association/dissociation rates	[87]
Isothermal Titration Calorimetry (ITC)	Measurement of heat changes during molecular binding	Quantification of binding affinity and stoichiometry	Binding enthalpy (ΔH), entropy (ΔS), affinity constants	[87]
Yeast Two-Hybrid (Y2H)	Protein-protein interaction detection via transcription activation	Identification of NBS-LRR interacting proteins in signaling complexes	Binary interaction confirmation, interaction networks	[3]
Molecular Docking	Computational prediction of protein-ligand binding poses	Analysis of NBS protein interactions with ADP/ATP and viral proteins	Binding energy scores, interaction interface residues	[3]

Computational Prediction Methods

Table 2: Computational Approaches for Interaction Prediction

Method Type	Examples	Key Features	Performance Metrics	Applications
Traditional Machine Learning	SVMrB, RotFB, RFB, C50B	Uses noncovalent interaction data (hydrogen bonding, aromatic interactions); 12 algorithms compared	Accuracy: ~0.70, Specificity: >0.92, Sensitivity: 0.35-0.68	Nanobody-ligand affinity prediction	[87]
Deep Learning	PRGminer (CNN-based)	Dipeptide composition input; two-phase prediction (R-gene identification and classification)	Accuracy: 98.75% (training), 95.72% (independent testing); MCC: 0.91	Plant resistance gene prediction and classification	[85]
Molecular Dynamics	GROMACS, HTMD	Simulation of protein dynamics and interaction stability	Binding free energy calculations, conformational stability	Nb–ligand complex stability assessment	[87]
Domain-Based Prediction	HMMER, InterProScan, PfamScan	Identification of NBS domains and architectural classification	Domain architecture patterns, orthogroup analysis	Genome-wide NBS gene identification	[3] [86]

Experimental Protocols for Key Methodologies

Machine Learning-Driven Affinity Prediction

Protocol for Nanobody-Ligand Affinity Prediction [87]:

• Data Collection and Processing:
  • Collect protein data bank (PDB) files of nanobody-ligand complexes from RCSB-PDB database
  • Calculate eight noncovalent interaction parameters using ProtInter tool:
    • Hydrophobic interactions (HI)
    • Disulfide bridges (DB)
    • Ionic interactions (IoInt)
    • Aromatic–aromatic interactions (AAI)
    • Aromatic–sulfur interactions (ASI)
    • Cation–pi interactions (CPI)
    • Hydrogen bonding (HBMM, HBMS, HBSS)
  • For each interaction type, calculate: count, mean, standard deviation, and quartile values of distances
  • Classify data into affinitive (MIC < 2000) and non-affinitive (MIC ≥ 2000) groups

• Model Training and Optimization:

  • Compare 12 machine learning algorithms including generalized linear model, naive Bayes, random forest, support vector machines, and multilayer perceptron
  • Apply 10-times 10-fold cross-validation with SMOTE sampling to address class imbalance
  • Remove features with correlation coefficients >0.7 to reduce dimensionality
  • Optimize hyperparameters for each algorithm based on caretr package guidelines
  • Evaluate models using accuracy, sensitivity, specificity, precision, F1 score, and Matthews correlation coefficient

• Model Validation:

  • Validate optimal models on independent test sets (20% of data)
  • Perform feature importance analysis to identify critical interactions
  • Stack uncorrelated models to improve prediction performance

Functional Validation via Virus-Induced Gene Silencing

Protocol for VIGS in Plant NBS-LRR Genes [5] [3]:

• Gene Selection and Vector Construction:
  • Identify target NBS-LRR genes through differential expression analysis or genome-wide association studies
  • Select 150-300 bp gene-specific fragments with low similarity to other genes
  • Clone fragments into appropriate VIGS vectors (TRV, BSMV, etc.)
  • Transform vectors into Agrobacterium tumefaciens for plant infiltration

• Plant Inoculation:

  • Grow plants to appropriate developmental stage (typically 2-4 leaf stage)
  • Infiltrate Agrobacterium carrying VIGS constructs into leaves using syringe or vacuum infiltration
  • Maintain plants under controlled conditions for systemic silencing establishment (2-3 weeks)

• Pathogen Challenge and Phenotyping:

  • Inoculate silenced plants with target pathogen at appropriate concentration
  • Maintain control plants (empty vector, non-silenced)
  • Monitor disease symptoms over time using standardized scoring systems
  • Quantify pathogen biomass through quantitative PCR or culture-based methods
  • Assess expression of silenced gene and defense markers using RT-qPCR

• Data Analysis:

  • Compare disease severity between silenced and control plants
  • Perform statistical analysis (ANOVA, t-tests) to determine significance
  • Correlated gene expression levels with phenotypic outcomes

The Scientist's Toolkit: Essential Research Reagents

Table 3: Key Research Reagents for NBS Interaction Studies

Reagent/Resource	Function	Example Applications	Sources
ProtInter	Computational tool for noncovalent interaction calculation from PDB files	Quantifying hydrogen bonding, aromatic interactions in Nb–ligand complexes	[87]
PRGminer	Deep learning-based resistance gene prediction and classification webserver	Identifying and classifying NBS-LRR genes in plant genomes	[85]
VIGS Vectors (TRV, BSMV)	Plant viral vectors for transient gene silencing	Functional validation of NBS-LRR genes in disease resistance	[5] [3]
PlantCARE Database	Cis-acting regulatory element prediction in promoter sequences	Identifying stress-responsive elements in NBS-LRR gene promoters	[50] [86]
Pfam/InterPro Databases	Protein domain identification and classification	Verifying NBS, CC, TIR, LRR domains in candidate proteins	[3] [86]
OrthoFinder	Phylogenetic orthology inference and comparative genomics	Identifying orthologous NBS-LRR gene groups across species	[3] [50]

Integrated Workflow for Comprehensive NBS Gene Analysis

This integrated workflow demonstrates the comprehensive approach required for thorough characterization of NBS domain genes, from initial identification to functional application. The synergy between computational predictions and experimental validations provides robust insights into protein-ligand and protein-protein interactions critical for plant immunity.

In the face of escalating biotic stresses that threaten global crop productivity, understanding the genetic basis of plant disease resistance has become a paramount research focus. A powerful approach in this endeavor involves comparative genomic analyses of susceptible and tolerant plant cultivars, which can reveal crucial resistance mechanisms and genetic elements. This guide synthesizes current research on the genomic differences between susceptible and tolerant cultivars, with a specific focus on the Nucleotide-Binding Site (NBS) domain genes—a major class of plant resistance genes. By objectively comparing performance across cultivars and presenting supporting experimental data, this review provides researchers and drug development professionals with a framework for identifying and utilizing genetic elements that confer disease resistance.

Performance Comparison: Susceptible vs. Tolerant Cultivars

Direct comparisons of susceptible and tolerant cultivars at the phenotypic, transcriptomic, and genomic levels reveal fundamental differences in their responses to pathogen attack.

Table 1: Comparative Performance of Susceptible and Tolerant Cultivars Under Pathogen Challenge

Cultivar/Condition	Pathogen	Key Performance Metrics	Major Findings	Citation
Gossypium hirsutum (Cotton)Tolerant: Mac7Susceptible: Coker 312	Cotton leaf curl virus (Begomovirus)	Genetic variation in NBS genes	Mac7 contained 6,583 unique variants in NBS genes, while Coker 312 contained 5,173.	[3]
Triticum aestivum (Wheat)Resistant: X413Susceptible: X73	Fusarium pseudograminearum (Fusarium crown rot)	Disease index, Hydrogen peroxide content, SOD activity	X413 exhibited stronger inhibition of fungal expansion, higher hydrogen peroxide content, and significantly higher SOD activity.	[88]
Lens ervoides (Lentil)Resistant: LR-66-629Susceptible: LR-66-570	Ascochyta lentis (Necrotroph)	Conidial germination, appressoria formation, necrotic area	The susceptible RIL had significantly higher conidial germination, more appressoria, and larger necrotic areas post-infection.	[89]
Triticum aestivum (Wheat)Resistant: ThatcherLr10Susceptible: Thatcher	Puccinia triticina (Leaf rust)	EST profiling, Gene expression	Resistant plants showed timely activation of defense genes (e.g., 14-3-3 protein, wali5 protein). Susceptible plants showed upregulation of senescence-associated genes.	[90]

The contrasting responses extend beyond visible symptoms to fundamental differences in gene expression networks. In wheat against Fusarium crown rot, the resistant germplasm X413 displayed fewer differentially expressed genes (DEGs) post-infection, which were notably enriched in resistance-related pathways like the lignin metabolic process and phenylpropanoid biosynthesis. In contrast, the susceptible X73 showed a greater number of DEGs, with significant downregulation of genes involved in growth and development, indicating a severe disruption of basic cellular processes upon pathogen challenge [88]. Similarly, in lentil against the necrotroph Ascochyta lentis, the resistant line demonstrated a stronger co-expression of genes involved in lipid localization, sulfur processes, and cellular responses to nutrients and stimuli [89].

Experimental Protocols for Comparative Genomics

To ensure the reproducibility of comparative studies, this section outlines the standard methodologies employed in the cited research.

Genomic DNA & RNA Sequencing Workflow

The foundational protocol for identifying genetic variants and expression differences involves a multi-step process from sample preparation to sequencing [3] [90] [88].

Genome-Wide Identification of NBS Domain Genes

A critical step is the comprehensive identification of NBS-LRR genes, which relies on domain-based searches [3] [46].

• Data Collection: Obtain the latest genome assemblies and proteome files for the target species from databases like NCBI, Phytozome, or Plaza.
• HMM Search: Use the PfamScan.pl script or HMMER to scan the proteome for the presence of the NB-ARC domain (Pfam: PF00931). A typical e-value cutoff is 1.1e-50 [3].
• Domain Architecture Analysis: Confirm the presence of the NBS domain and identify associated domains (e.g., TIR, CC, LRR) using tools like Pfam and the Conserved Domain Database (CDD) [46].
• Classification: Classify the identified genes based on their domain architecture (e.g., TNL, CNL, RNL, or atypical NBS) [3] [46].

Orthogroup and Evolutionary Analysis

To trace the evolution of NBS genes across species, orthologs are identified and grouped [3].

• Sequence Clustering: Input protein sequences of identified NBS genes from multiple species into OrthoFinder v2.5.1, which uses DIAMOND for sequence similarity and MCL for clustering [3].
• Phylogenetic Reconstruction: Perform multiple sequence alignment of orthogroups (OGs) using MAFFT 7.0. Construct a maximum likelihood phylogenetic tree using tools like FastTreeMP with bootstrapping (e.g., 1000 replicates) [3].
• Duplication Analysis: Identify tandem and segmental duplication events within the genome to understand the mechanisms of gene family expansion [3] [91].

Expression Profiling and Functional Validation

Linking genetic data to function requires transcriptomic and functional validation [3] [88] [89].

• Expression Data Retrieval/Analysis: Retrieve FPKM/TPM values from RNA-seq databases or process raw RNA-seq data. Categorize expression into tissue-specific, abiotic stress-specific, and biotic stress-specific profiles [3].
• Differential Expression & Co-expression Analysis: Identify DEGs between resistant and susceptible lines post-inoculation using tools like DESeq2. Perform Weighted Gene Co-expression Network Analysis (WGCNA) to identify hub genes and key modules associated with resistance [88] [89].
• Functional Validation via VIGS: Validate the role of candidate NBS genes using Virus-Induced Gene Silencing (VIGS). Design a TRV-based vector containing a fragment of the target gene, agro-infiltrate it into resistant plants, and then challenge with the pathogen. Assess disease severity and viral titer compared to control plants [3].

The Scientist's Toolkit: Key Research Reagents & Solutions

Table 2: Essential Reagents and Tools for Comparative Genomics of Disease Resistance

Reagent/Tool Name	Function in Research	Application Context
HMMER (PfamScan)	Identifies protein domains (e.g., NB-ARC) using hidden Markov models.	Genome-wide identification and classification of NBS-LRR genes [3] [46].
OrthoFinder	Infers orthogroups and gene families across multiple species.	Evolutionary analysis, identification of core and species-specific orthogroups [3].
DESeq2 / EdgeR	Statistical analysis of differential gene expression from RNA-seq data.	Identifying genes upregulated in tolerant cultivars post-pathogen challenge [88].
WGCNA	Constructs gene co-expression networks to identify hub genes and functional modules.	Uncovering key regulatory networks and genes associated with resistance traits [88] [89].
VIGS Vectors (e.g., TRV)	Mediates transient gene silencing for rapid functional validation.	Testing the requirement of candidate NBS genes for resistance [3].
qPCR Reagents	Quantifies gene expression and pathogen biomass accurately.	Validating RNA-seq results and tracking pathogen growth in hosts [89].

Signaling Pathways in Plant Immunity

The defense response in plants is a multi-layered system. The following diagram integrates the key pathways and components involved, particularly highlighting the role of NBS-LRR proteins in Effector-Triggered Immunity (ETI).

The model shows that surface-localized Pattern Recognition Receptors (PRRs) activate Pattern-Triggered Immunity (PTI) upon recognition of Pathogen-Associated Molecular Patterns (PAMPs) [32] [92]. Adapted pathogens secrete effector proteins to suppress PTI. In resistant plants, intracellular NBS-LRR (NLR) proteins directly or indirectly recognize these effectors, activating a stronger immune response known as Effector-Triggered Immunity (ETI) [32] [46]. ETI is often associated with a Hypersensitive Response (HR), a form of programmed cell death at the infection site that restricts pathogen spread [32] [89]. Recent studies show that PTI and ETI synergistically amplify each other, leading to a robust defense output [32] [46]. This includes systemic acquired resistance (SAR), reinforcement of cell walls via lignin deposition, and production of antimicrobial pathogenesis-related (PR) proteins [90] [88].

Comparative genomics of susceptible and tolerant cultivars provides an powerful strategy to unravel the complex genetic architecture of disease resistance. The consistent trend across studies is that resistant genotypes possess a more efficient surveillance and response system, often characterized by specific NBS-LRR gene variants, a timely and balanced transcriptomic reprogramming favoring defense over growth, and the activation of key biochemical pathways like phenylpropanoid biosynthesis. The experimental frameworks and tools detailed in this guide offer a roadmap for researchers to identify and validate critical resistance genes. The integration of these genomic insights into breeding programs, including through modern techniques like CRISPR-mediated genome editing [92], holds great promise for developing durable disease-resistant crops, which is essential for future global food security.

Within the framework of a broader comparative analysis of Nucleotide-Binding Site (NBS) domain genes across plant species, this case study investigates the pivotal role of specific NBS-Leucine-Rich Repeat (NBS-LRR) genes in conferring resistance to Fusarium wilt in tung trees. Fusarium wilt, caused by soil-borne pathogens of the Fusarium genus, represents a significant threat to global agriculture, affecting a wide range of staple and economically important crops, resulting in substantial yield losses and economic impacts [93]. Plants have evolved a sophisticated immune system where the NBS-LRR gene family, encoding the largest class of intracellular resistance (R) proteins, plays a critical role in effector-triggered immunity (ETI) [46]. This analysis of tung tree (Vernicia species) offers a unique comparative model between a resistant and a susceptible genotype, providing insights that are applicable to disease resistance breeding in other plant species.

Comparative Genomic Analysis of NBS-LRR Genes

Genomic Distribution and Diversity in Tung Trees

A genome-wide identification in the two principal tung tree cultivars, the susceptible Vernicia fordii and the resistant Vernicia montana, revealed a total of 239 NBS-LRR genes—90 in V. fordii and 149 in V. montana [48] [94] [5]. This discrepancy in gene number immediately suggests a potential genomic basis for the difference in disease resistance.

• Gene Classification: The NBS-LRR genes were categorized based on their N-terminal domains and the presence of the C-terminal LRR domain.
• Domain Analysis: In V. fordii, NBS-LRRs fell into four subgroups, with 54.4% containing a Coiled-Coil (CC) domain. Notably, no TIR-NBS-LRR (TNL) types were identified [48] [5].
• V. montana's NBS-LRRs were more diverse, classified into seven subgroups. Among them, 65.8% contained CC domains, and 8.1% possessed TIR domains, including two genes with both CC and TIR domains [48] [5].
• LRR Domain Variation: The LRR domain, crucial for pathogen recognition, was more varied in V. montana, which possessed four types of LRR domains compared to only two in V. fordii. The absence of LRR1 and LRR4 domains in V. fordii indicates evolutionary loss events that may contribute to its susceptibility [48].

Cross-Species Comparison of NBS-LRR Genes

The NBS-LRR gene family is a cornerstone of plant innate immunity, and its characteristics vary significantly across species, as summarized in Table 1. These variations in family size, composition, and genomic organization reflect the dynamic evolution of this gene family in response to pathogen pressures [79].

Table 1: Comparative Analysis of NBS-LRR Genes across Select Plant Species

Plant Species	Total NBS-LRR Genes	TNL Genes	CNL Genes	Key Genomic Features
Vernicia fordii (Tung tree)	90	0	49 (CC-containing)	Absence of TNL genes; fewer LRR domain types [48].
Vernicia montana (Tung tree)	149	12 (TIR-containing)	98 (CC-containing)	Presence of both TNL and CNL types; diverse LRR domains [48].
Arabidopsis thaliana	149-159	94-98	55	Model dicot with a balanced ratio; genes are unevenly distributed on chromosomes [79].
Oryza sativa (Rice)	553-653	0	553-653 (approx.)	A monocot; complete absence of TNL genes; one of the largest NBS-LRR families [79].
Solanum tuberosum (Potato)	435-438	65-77	361-370	High number of CNL genes; enriched on specific chromosomes (e.g., Chr4, Chr11) [79].
Salvia miltiorrhiza	196	2	75 (CC-containing)	Medicinal plant with a marked reduction in TNL and RNL subfamilies [46].
Nicotiana benthamiana	156	5 (TNL-type)	25 (CNL-type)	Model plant for virology; includes various truncated forms (N-type, CN-type, etc.) [71].

A key evolutionary pattern observed across species is the differential presence of the TNL subclass. While TNL genes are common in dicots like Arabidopsis thaliana, they are completely absent in monocots such as rice, wheat, and maize [79] [46]. The loss of TNL genes in a eudicot like V. fordii is a relatively rare event, previously reported only in a few species like Sesamum indicum, and highlights the dynamic nature of R gene evolution [48] [5]. Furthermore, NBS-LRR genes are typically distributed non-randomly across chromosomes, often forming clusters that facilitate the rapid evolution of new resistance specificities through mechanisms like tandem duplication and unequal crossing-over [48] [79] [14].

Experimental Characterization of a Key Resistance Gene

Identification of a Candidate Fusarium Wilt Resistance Gene

A comparative transcriptomic analysis between V. fordii and V. montana following Fusarium wilt infection pinpointed a critical orthologous gene pair: Vf11G0978 in V. fordii and Vm019719 in V. montana [48] [94] [5]. This pair exhibited starkly contrasting expression patterns:

• Vm019719 showed upregulated expression in the resistant V. montana.
• Vf11G0978 showed downregulated expression in the susceptible V. fordii [48] [94].

This inverse correlation strongly suggested that Vm019719 was a candidate gene for mediating Fusarium wilt resistance in V. montana.

Functional Validation via Virus-Induced Gene Silencing (VIGS)

To confirm the function of Vm019719, researchers employed Virus-Induced Gene Silencing (VIGS), a powerful tool for rapid functional genomics in plants [48] [71].

Experimental Protocol:

• Vector Construction: A fragment of the Vm019719 gene was cloned into a TRV-based VIGS vector.
• Plant Material: Seedlings of resistant V. montana were selected for the experiment.
• Agroinfiltration: The recombinant VIGS vector was introduced into Agrobacterium tumefaciens, which was then infiltrated into the leaves of V. montana seedlings. Control plants were infiltrated with an empty vector.
• Pathogen Challenge: After giving time for the VIGS system to silence the target gene, plants were inoculated with Fusarium oxysporum.
• Phenotypic Assessment: Disease symptoms and plant health were monitored and recorded over time.
• Molecular Verification: Silencing of Vm019719 was confirmed using quantitative real-time PCR (qRT-PCR) to measure transcript levels [48].

Result: The VIGS experiment provided direct functional evidence. V. montana plants with silenced Vm019719 expression lost their resistance and showed increased susceptibility to Fusarium wilt, comparable to the phenotype of V. fordii [48] [94]. This confirmed that Vm019719 is essential for resistance in V. montana.

Elucidating the Molecular Basis of Susceptibility

Further investigation revealed the precise regulatory mechanism behind the differential expression of this orthologous gene pair.

• Promoter Analysis: The promoter region of the resistant Vm019719 allele in V. montana contains a functional W-box cis-element.
• Transcriptional Activation: This W-box is recognized and bound by the transcription factor VmWRKY64, which activates the expression of Vm019719, leading to an effective defense response [48].
• Promoter Defect in Susceptible Allele: In the susceptible V. fordii allele Vf11G0978, a deletion in the promoter region has resulted in the loss of this critical W-box element [48] [94]. Consequently, the gene cannot be properly activated in response to pathogen attack, leading to an ineffective defense and susceptibility.

This mechanism can be visualized in the following pathway diagram.

The Scientist's Toolkit: Key Research Reagent Solutions

The functional characterization of NBS-LRR genes relies on a suite of specialized reagents and methodologies. The following table details essential tools for research in this field, as exemplified by the tung tree case study.

Table 2: Essential Research Reagents and Methods for NBS-LRR Gene Analysis

Research Tool / Reagent	Function & Application in NBS-LRR Research
HMMER Software	Function: Bioinformatics tool for sequence analysis. Application: Genome-wide identification of NBS-LRR genes using hidden Markov models (HMMs) based on the conserved NBS (NB-ARC) domain (PF00931) [48] [46] [71].
VIGS Vectors (e.g., TRV)	Function: Knocks down gene expression without generating stable transgenic lines. Application: Rapid functional validation of candidate R genes (e.g., Vm019719) by silencing them and assessing changes in disease phenotype [48] [94].
qRT-PCR Assays	Function: Precisely quantifies gene expression levels. Application: Measures the expression dynamics of NBS-LRR genes in response to pathogen infection and verifies the efficiency of VIGS [48].
MEME Suite	Function: Discovers conserved motifs in protein or DNA sequences. Application: Identifies and visualizes conserved structural motifs within NBS-LRR protein sequences, aiding in phylogenetic classification [71].
Phylogenetic Analysis Tools (e.g., MEGA)	Function: Infers evolutionary relationships. Application: Classifies NBS-LRR genes into subfamilies (TNL, CNL, RNL) and identifies orthologs and paralogs across species [46] [71].
CRISPR-Cas Systems	Function: Targeted genome editing. Application: Knocks out susceptibility (S) genes or precisely introduces specific R genes (like Vm019719) into susceptible cultivars to enhance resistance [93].

The experimental workflow for identifying and validating a candidate NBS-LRR gene, from genomic analysis to functional characterization, is summarized below.

This case study demonstrates that the functional NBS-LRR gene Vm019719, activated by the transcription factor VmWRKY64, is a key determinant of Fusarium wilt resistance in Vernicia montana. The susceptibility of its cultivated relative, V. fordii, is directly linked to a promoter deletion that disrupts this regulatory circuit [48] [94]. This finding provides a clear target for molecular breeding.

Looking forward, the application of CRISPR-Cas genome editing technology presents a promising avenue for directly improving disease resistance [93]. Strategies could include:

• Knocking out susceptibility (S) genes in susceptible crops.
• Precisely editing the promoter regions of allelic R genes in susceptible varieties, such as engineering a W-box into the Vf11G0978 promoter in V. fordii, to restore their responsiveness.
• Multiplex editing to pyramid multiple R genes, creating durable and broad-spectrum resistance [93].

The comparative analysis of NBS-LRR genes across plant species underscores their fundamental role in plant immunity while highlighting the species-specific innovations that have evolved to combat pathogens. The knowledge gained from model systems like tung tree provides a powerful toolkit for engineering disease-resistant crops, thereby contributing to global food security.

Conclusion

The comparative analysis of NBS domain genes reveals a dynamic evolutionary landscape shaped by duplication events and selective pressures, resulting in vast diversity essential for plant adaptation. Key takeaways include the central role of tandem duplications in species-specific resistance gene expansion, the power of integrated computational and functional genomics for gene discovery, and the critical link between specific NBS-LRR variants and disease resistance phenotypes in crops. For biomedical and clinical research, the sophisticated mechanisms of plant immune receptors—particularly their modular domain architecture and specific protein-ligand interactions—offer inspirational blueprints for designing novel therapeutic proteins and molecular scaffolds. Future research should focus on harnessing machine learning for predicting resistance specificity, engineering NBS genes for broad-spectrum disease resistance in crops, and exploring the potential of plant-derived resistance protein architectures for developing new diagnostic and therapeutic agents in medicine.

References

• 1. Distinct Evolutionary Patterns of NBS-Encoding Genes in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7365912/]
• 2. Multilevel evolution shapes the function of NB-LRR ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.1007288/full]
• 3. Comparative analysis, diversification, and functional ... [https://www.nature.com/articles/s41598-024-62876-5]
• 4. Comparative analysis of plant immune receptor architectures ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC4759884/]
• 5. Functional characterization of NBS-LRR genes reveals an ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10898138/]
• 6. Identification and Characterization of NBS Resistance ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2021.758559/full]
• 7. Evolution patterns of NBS genes in the genus Dendrobium ... [https://bmcplantbiol.biomedcentral.com/articles/10.1186/s12870-022-03904-2]
• 8. Genome-wide comparative analysis of NBS-encoding genes ... [https://bmcgenomics.biomedcentral.com/articles/10.1186/1471-2164-15-3]
• 9. Genome-wide analysis of NBS-LRR genes in Rosaceae ... [https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2022.1052191/full]
• 10. miR825‐5p‐regulated TNLs govern Arabidopsis resistance to ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12371183/]
• 11. Structure/Function Approaches to Understanding CC Domain ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6290485/]
• 12. Comparative Analysis of NBS Domain Sequences ... [https://pubmed.ncbi.nlm.nih.gov/15019616/]
• 13. A compendium of genome-wide sequence reads from NBS ... [https://www.nature.com/articles/s41598-020-67848-z]
• 14. Plant NBS-LRR proteins: adaptable guards - Genome Biology [https://genomebiology.biomedcentral.com/articles/10.1186/gb-2006-7-4-212]
• 15. Large-Scale Analyses of Angiosperm Nucleotide-Binding Site ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC4825152/]
• 16. A Primary Survey on Bryophyte Species Reveals Two Novel ... [https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0036700]
• 17. A Primary Survey on Bryophyte Species Reveals Two ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC3352924/]
• 18. Bryophytes hold a larger gene family space than vascular ... [https://www.nature.com/articles/s41588-025-02325-9]
• 19. Comparative analysis of the NLR gene family in ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2025.1681919/full]
• 20. Comparative analysis of NBS domain sequences ... [https://www.sciencedirect.com/science/article/abs/pii/S1055790303002744]
• 21. The Diversification of Plant NBS-LRR Defense Genes Directs ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5026261/]
• 22. Interaction between domains of a plant NBS–LRR protein in ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC126192/]
• 23. Gene duplication and evolution in recurring polyploidization ... [https://genomebiology.biomedcentral.com/articles/10.1186/s13059-019-1650-2]
• 24. Importance of Lineage-Specific Expansion of Plant ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC2556807/]
• 25. The impact of tandem duplication on gene evolution in ... [https://www.sciencedirect.com/science/article/pii/S2095311921636985]
• 26. Spatial transcriptomics illuminates evolutionary fates - PMC [https://pmc.ncbi.nlm.nih.gov/articles/PMC12622374/]
• 27. Genome-wide comparative analysis of NBS-encoding genes ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5388996/]
• 28. Evolutionary balance between LRR domain loss and young ... [https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-019-6212-1]
• 29. Comparative analysis of NBS-LRR genes and their ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5291535/]
• 30. Pangenome analysis reveals structural variation ... [https://www.nature.com/articles/s41588-025-02170-w]
• 31. Systematic Analysis and Comparison of Nucleotide-Binding ... [https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0068435]
• 32. Computational Strategies for Elucidating Plant Disease ... [https://www.sciencedirect.com/science/article/abs/pii/S0885576525003790]
• 33. Evolutionary dynamics of protein domain architecture in plants [https://pmc.ncbi.nlm.nih.gov/articles/PMC3310802/]
• 34. OrthoFinder: phylogenetic orthology inference for comparative ... [https://genomebiology.biomedcentral.com/articles/10.1186/s13059-019-1832-y]
• 35. Plant genome and transcriptome annotations [https://pmc.ncbi.nlm.nih.gov/articles/PMC5952960/]
• 36. New Tools in Orthology Analysis: A Brief Review of Promising ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC5674930/]
• 37. Comprehensive evolutionary analysis and nomenclature of ... [https://www.life-science-alliance.org/content/5/9/e202101328]
• 38. PRGminer: harnessing deep learning for the prediction of ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2025.1411525/full]
• 39. Identification, characterization, and validation of NBS ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10318170/]
• 40. Evolutionary analysis of TIR- and non-TIR-NBS-LRR ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2024.1452251/full]
• 41. DPFunc: accurately predicting protein function via deep ... [https://www.nature.com/articles/s41467-024-54816-8]
• 42. Classification and prediction of protein – protein interaction interface... [https://www.nature.com/articles/s41598-020-80900-2?error=cookies_not_supported&code=6a5f0c5b-dd1c-421a-92a3-4786a879876c]
• 43. Protein structure prediction via deep learning: an in-depth ... [https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2025.1498662/full]
• 44. GitHub - rdk/p2rank: P2Rank: Protein -ligand binding site prediction ... [https://github.com/rdk/p2rank]
• 45. Recent advances in deep learning for protein-protein interaction [https://pmc.ncbi.nlm.nih.gov/articles/PMC12168265/]
• 46. Genome-Wide Characterization and Expression Pattern ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12469828/]
• 47. sequentiabiotech/DRAGO-API: Disease Resistance ... [https://github.com/sequentiabiotech/DRAGO-API]
• 48. Functional characterization of NBS-LRR genes reveals an ... [https://bmcbiol.biomedcentral.com/articles/10.1186/s12915-024-01836-x]
• 49. Discovery of functional NLRs using expression level, high- ... [https://www.nature.com/articles/s41477-025-02110-w]
• 50. Comparative analysis of the NLR gene family in the genomes ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12507765/]
• 51. Genome-Wide Identification of NBS-LRR Family in Three ... [https://www.mdpi.com/2073-4425/16/6/680]
• 52. PlantTribes2: Tools for comparative gene family analysis in ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.1011199/full]
• 53. Assembling genomes of non‐model plants: A case study with ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12448783/]
• 54. Establishing genome sequencing and assembly for non ... [https://frontiersinzoology.biomedcentral.com/articles/10.1186/s12983-025-00561-7]
• 55. Transcriptome profiling and RNA-Seq SNP analysis of ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2025.1532943/full]
• 56. Identification and expression profiling analysis of NBS ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC6500516/]
• 57. Genome-Wide Characterization and Expression Pattern ... [https://www.mdpi.com/1422-0067/26/18/9063]
• 58. Striking convergent selection history of wheat and barley ... [https://www.nature.com/articles/s41477-025-02128-0]
• 59. Comparative single-cell transcriptomic map reveals ... [https://www.sciencedirect.com/science/article/pii/S2090123225002565]
• 60. Integrating genetic and gene expression data in network ... [https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-025-06143-y]
• 61. Plant NBS-LRR proteins: adaptable guards - PubMed Central [https://pmc.ncbi.nlm.nih.gov/articles/PMC1557992/]
• 62. Genome-Wide Association Study Identifies NBS-LRR ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2017.01398/full]
• 63. Identification and distribution of the NBS-LRR gene family in ... [https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-015-1554-9]
• 64. Plant GARDEN: a portal website for cross-searching between ... [https://bmcplantbiol.biomedcentral.com/articles/10.1186/s12870-023-04392-8]
• 65. A Compendium of Sequenced NBS-domains of Resistance ... [https://db.cngb.org/data_resources/project/PRJEB83917]
• 66. Phylogenetic, Structural, and Evolutionary Insights into ... [https://www.mdpi.com/1422-0067/26/5/1828]
• 67. Genetic Divergence of Lineage-Specific Tandemly ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2022.875202/full]
• 68. Diversity, abundance, and domain architecture of plant ... [https://www.sciencedirect.com/science/article/pii/S2405844024105063]
• 69. Conservation and Tandem Duplication of tRNA Genes in ... [https://www.mdpi.com/2073-4425/16/11/1307]
• 70. Genome-Wide Identification of NBS-LRR Family in Three ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12192410/]
• 71. Genome-wide identification and characterization of NBS ... [https://www.nature.com/articles/s41598-025-03507-5]
• 72. A long road ahead to reliable and complete medicinal plant ... [https://www.nature.com/articles/s41467-025-57448-8]
• 73. Best Practices for Data Quality in Machine Learning [https://www.anomalo.com/blog/data-quality-in-machine-learning-best-practices-and-techniques/]
• 74. Class-imbalanced datasets | Machine Learning [https://developers.google.com/machine-learning/crash-course/overfitting/imbalanced-datasets]
• 75. Handling Imbalanced Data for Classification [https://www.geeksforgeeks.org/machine-learning/handling-imbalanced-data-for-classification/]
• 76. Enhancing Data Quality Using Machine Learning [https://medium.com/@melissadunkeld/enhancing-data-quality-using-machine-learning-4003d4f1fbdb]
• 77. Improving Data Quality Using AI and ML [https://www.dataversity.net/articles/improving-data-quality-using-ai-and-ml/]
• 78. The Role of AI and Machine Learning in Automating Data ... [https://firsteigen.com/blog/the-role-of-ml-and-ai-in-data-quality-management/]
• 79. Plant Nucleotide Binding Site–Leucine-Rich Repeat (NBS ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC3645687/]
• 80. Implication of the LRR Domain in the Regulation and ... [https://www.mdpi.com/2073-4409/13/16/1365]
• 81. Comparative-genomic analysis reveals dynamic NLR gene ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10017999/]
• 82. Virus-Induced Gene Silencing (VIGS) in Functional Genomics [https://www.mdpi.com/2311-7524/11/11/1297]
• 83. Transcriptome Analysis and Functional Characterization of the ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12348818/]
• 84. Efficient Virus-Induced Gene Silencing (VIGS) Method for ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12115190/]
• 85. PRGminer: harnessing deep learning for the prediction of ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC12170542/]
• 86. Comparative genomics and bioinformatics approaches ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC10929019/]
• 87. Machine Learning-Driven Methods for Nanobody Affinity ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11618429/]
• 88. Phenotypic and comparative transcriptomic analyses of ... [https://bmcgenomics.biomedcentral.com/articles/10.1186/s12864-025-12237-x]
• 89. Interactive Gene Expression Patterns of Susceptible and ... [https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2020.01259/full]
• 90. Comparative Gene Expression Analysis of Susceptible and ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC2920755/]
• 91. Comparative genomics and synteny analysis of PP2C ... [https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2025.1596800/full]
• 92. CRISPR-mediated genome editing of wheat for enhancing ... [https://pmc.ncbi.nlm.nih.gov/articles/PMC11893844/]
• 93. The Enhancement of Fungal Disease Resistance in Major ... [https://www.mdpi.com/2073-4425/16/11/1263]
• 94. Functional characterization of NBS-LRR genes reveals an NBS-LRR gene that mediates resistance against Fusarium wilt - PubMed [https://pubmed.ncbi.nlm.nih.gov/38408951/]