How AI Is Improving Chickpea Genetics and Crop Development
- The global chickpea market reached USD 15.96 billion in 2024 and is projected to grow at a CAGR of 7.1% through 2033 (SkyQuest, 2025), yet production has remained nearly stagnant per hectare for decades — a gap that artificial intelligence is now closing at remarkable speed.
- AI helps design the perfect chickpea by analyzing more than 3,000 cultivated and wild varieties, identifying 1,582 novel genes, and predicting breeding outcomes that would take conventional programs a generation to achieve.
- From drought tolerance and disease resistance to heat adaptation and higher protein content, AI-driven genomic selection platforms are rewriting the rules of crop development.
The global chickpea market reached USD 15.96 billion in 2024 and is growing at a CAGR of 7.1% toward 2033. Against this backdrop of surging demand, a critical question faces breeders, agronomists, and food system planners: can crop science keep up? AI helps design the perfect chickpea not as a single ideal plant, but as a precisely optimized variety engineered for a specific environment, climate, and nutritional purpose — which is exactly what modern agriculture needs.
What Is the Perfect Chickpea?
Chickpeas, known botanically as Cicer arietinum, rank third among all pulses in global production and are cultivated in more than 50 countries, with South Asia and sub-Saharan Africa leading in both area and consumption.
For hundreds of millions of people in these regions, chickpeas are not a dietary choice — they are a dietary cornerstone, providing affordable plant-based protein where animal protein is either scarce or unaffordable. A chickpea variety that yields more, survives drought, and resists disease does not just benefit farmers economically; it shores up food security at a population scale.
The role of AI in modern crop development has grown from a research curiosity into a practical, deployable toolkit. Machine learning models can now process genomic datasets containing millions of genetic markers and return trait predictions in hours — a task that would require decades of field observation under traditional breeding frameworks.
Why Chickpea Improvement Matters More Than Ever
The plant-based protein movement is not a passing trend. As consumers in North America, Europe, and parts of Asia continue reducing meat consumption, demand for legume-based protein sources has intensified sharply. Chickpeas, which contain approximately 19–20 grams of protein per 100 grams of dry weight, are among the most efficient protein sources available at scale.
The challenge is that global pulse productivity has remained largely stagnant for five decades, creating a growing gap between demand and supply that conventional breeding has struggled to close.
Climate change is making this challenge worse, not better. Chickpea is a cool-season legume, and rising temperatures in its traditional growing regions — particularly northern India, Pakistan, Ethiopia, and Australia — are compressing the planting window and triggering heat stress during flowering, which is the most yield-sensitive stage of the crop’s life cycle.
Simultaneously, rainfall patterns in semi-arid regions are becoming less predictable, increasing the frequency of both drought events and out-of-season wet spells that promote fungal disease.
i. Fusarium wilt (a soilborne fungal disease that can destroy entire fields in infected plots) and Ascochyta blight (a foliar fungal disease causing up to 100% crop loss in severe epidemics) remain the two most economically damaging diseases in chickpea production worldwide, and developing resistant varieties through traditional crossing is slow and resource-intensive.
ii. Water scarcity is intensifying across South Asia and sub-Saharan Africa, where chickpea yields are already suppressed well below their genetic potential. Breeding for deeper root architecture, stomatal efficiency, and osmotic adjustment has become a strategic priority for every major agricultural research institute working with the crop.
iii. Nutritional improvement is increasingly targeted by breeders, as biofortification (the process of increasing the concentration of vitamins and minerals in crops through breeding) offers a cost-effective alternative to food fortification programs in low-income countries.
Each of these challenges requires simultaneous optimization of multiple traits — a complexity that makes traditional breeding approaches painfully slow and AI-based approaches genuinely transformative.
Traditional Chickpea Breeding Versus AI-Assisted Breeding
Conventional plant breeding works through a cyclical process of crossing two parent lines with desirable complementary traits, then selecting offspring over multiple generations for the combination of those traits. This process, known as phenotypic selection (choosing plants based on their visible characteristics rather than their genetic code), is reliable but slow.
A single cycle of crossing, evaluation, and stabilization in chickpea takes between eight and twelve years, and breeders can only evaluate a limited number of plants per season due to field space and labor constraints.
The limitations of this approach become clear when you consider the mathematics. A chickpea plant carries approximately 738 million base pairs of DNA spread across eight chromosomes, and the traits breeders care about — yield, drought tolerance, disease resistance — are each controlled by dozens or hundreds of individual genes interacting in complex networks.
Selecting for all of these simultaneously using field observation alone is practically impossible within a commercially useful timeframe. AI-assisted breeding disrupts this bottleneck at three levels.
- First, it enables genomic selection (predicting a plant’s breeding value directly from its DNA sequence, without needing to grow the plant to maturity), which collapses the evaluation cycle from years to weeks.
- Second, machine learning models trained on large genetic datasets can identify non-obvious relationships between genetic markers and traits that human experts would never find through manual analysis.
- Third, AI platforms can simulate thousands of crossing combinations and predict the performance of each hypothetical offspring, allowing breeders to prioritize the most promising crosses before committing a single seed to soil.
The time and cost implications are significant. Research programs using AI-assisted breeding have reported cycle time reductions of 30–50% compared to conventional approaches, with corresponding reductions in the number of field trials required to identify superior lines.
How AI Analyzes Chickpea Genetics at the Molecular Level
The foundation of AI-assisted chickpea breeding is genomic data — specifically, the identification of single nucleotide polymorphisms (SNPs), which are single-letter differences in the DNA sequence between individual plants. SNPs serve as genetic markers: flags planted at known positions in the genome that indicate which version of a nearby gene a plant carries.
A chickpea genome-wide association study (GWAS) typically analyzes hundreds of thousands to millions of SNPs simultaneously across hundreds or thousands of plant varieties, looking for SNPs that are statistically associated with a trait of interest.
In 2021, an international team led by researchers at the University of Queensland and ICRISAT sequenced the genomes of more than 3,366 chickpea varieties, including both cultivated types and wild relatives.
This dataset, the largest chickpea pan-genome ever assembled, identified 1,582 novel genes not previously catalogued and provided the raw material for AI models to learn the genetic architecture of complex agronomic traits.
Varshney et al. (Nature, 2021) found that haplotype-based genomic prediction models increased predicted seed weight performance by up to 23% in chickpea populations compared to conventional breeding approaches, while OCS-based pre-breeding strategies maintained genetic diversity across all 16 traits evaluated.
Chickpea breeders can now use AI-generated haplotype maps to design crosses that maximize yield-related traits without narrowing the genetic base — a critical safeguard against long-term vulnerability.
Machine learning models built on this genomic data operate by learning statistical patterns — specifically, which combinations of SNP alleles (the alternative versions of a gene at a specific position) reliably predict a particular trait score. Once trained, these models function as predictive engines:
- feed in the SNP profile of a new,
- uncharacterized plant, and
- the model returns predicted scores for yield potential, drought tolerance, disease resistance, and other traits. The better the training data, the more accurate the predictions.
Key Traits AI Helps Optimize in Chickpeas
1. Higher Yield Through Genomic Intelligence
Yield in chickpea is a composite trait influenced by seed size, seeds per pod, pods per plant, days to flowering, and above-ground biomass, all of which are themselves influenced by dozens of genes each.
AI approaches this complexity by treating yield as a systems-level outcome rather than a single target, modeling the interactions among contributing traits simultaneously rather than optimizing them one at a time. The University of Queensland’s FastStack technology platform combines AI with genomic prediction to identify gene combinations most likely to improve overall crop performance.
Using FastStack on the 3,366-variety dataset, researchers modeled an “ideal” chickpea and predicted yield increases of up to 12% for seed weight, a key yield proxy, compared to the best currently available commercial varieties. This 12% figure represents the outcome of selecting from a virtually unlimited pool of crossing combinations — something no human breeder could evaluate manually.
2. Drought Tolerance Through Genetic Mapping
Breeding for drought tolerance requires understanding a cascade of biological responses: root depth and architecture, leaf area regulation, stomatal conductance (the rate at which leaves open and close pores to manage water loss), osmotic adjustment (the plant’s ability to maintain cell function under low water availability), and reproductive stage resilience.
AI models can simultaneously track SNPs associated with each of these sub-traits and identify plant lines that score strongly across all of them.
ICRISAT researchers have identified specific QTLs (quantitative trait loci — genomic regions that explain measurable variation in a trait) linked to root depth in chickpea under terminal drought conditions, providing AI models with the genetic coordinates for targeted selection.
Deep learning models trained on multi-environment trial data — datasets where the same varieties are grown across many locations with different rainfall patterns — can predict drought performance in a new target environment with substantially greater accuracy than single-location field data alone.
3. Disease Resistance Against Fusarium Wilt and Ascochyta Blight
Disease resistance is one of the most technically challenging traits to breed for, because pathogen populations evolve rapidly and can overcome newly developed resistance genes within a few crop seasons.
AI approaches this problem not by targeting single resistance genes but by stacking multiple resistance mechanisms — a strategy called pyramiding — to create varieties that are far harder for pathogens to overcome.
Lin et al. (The Plant Genome, 2025) assessed genomic selection for Ascochyta blight resistance in chickpea breeding germplasm in Australia, finding that genomic selection models achieved prediction accuracies of 0.50–0.65 for blight resistance scores, demonstrating that AI-based pre-selection can significantly reduce the number of field disease nurseries required per breeding cycle.
Reducing field disease nurseries through genomic pre-selection cuts the time and cost of resistance breeding while identifying durable stacking combinations that no single-gene approach could achieve.
Genome-wide association studies have identified multiple independent genomic regions linked to Ascochyta blight resistance in Australian chickpea germplasm. AI models trained on this data can evaluate thousands of breeding lines for predicted resistance levels, flagging the most robust candidates for field confirmation — which concentrates experimental resources where they are most likely to produce results.
4. Heat Tolerance for a Warming Climate
A 2024 GWAS study by Jeffrey et al. (Frontiers in Plant Science, 2024) identified QTLs for heat tolerance linked specifically to canopy closure timing and early flowering in chickpea — two traits that allow the crop to complete its reproductive stage before peak summer temperatures arrive.
The National Institute of Plant Genome Research (NIPGR) in New Delhi used a next-generation combinatorial genomic strategy to identify 10 key genomic regions associated with heat stress tolerance traits in chickpea, naming specific regulatory genes including RAD23b, CIPK25, and WRKY40 as priority targets for AI-guided introgression (Plant, Cell and Environment, 2025).
These genomic coordinates give AI prediction models the anchors they need to screen large germplasm collections for heat-adapted individuals, dramatically narrowing the breeding population that needs to be evaluated in hot-environment field trials.
5. Improved Nutritional Value Through Biofortification
Protein content, iron bioavailability, and zinc concentration in chickpea seeds vary significantly across varieties, and the genetic basis of these differences is now being mapped with increasing precision.
AI models trained on metabolomics data — measurements of the full suite of chemical compounds present in a seed sample — can predict nutritional profiles from genomic data, enabling simultaneous selection for agronomic and nutritional traits within the same breeding pipeline.
This is particularly relevant for Sub-Saharan Africa, where chickpea consumption is rising rapidly and iron deficiency anemia remains a major public health burden. Breeding chickpea varieties with 20–30% higher iron bioavailability could deliver measurable population-level health impacts without requiring any change in dietary habits.
AI Technologies Powering Chickpea Research
The term “AI” encompasses several distinct technologies, each contributing differently to chickpea breeding programs. Understanding which tools do what helps practitioners assess which investments make sense for their specific programs.
- Machine learning (algorithms that learn predictive patterns from labeled training data) is used primarily for genomic selection — predicting breeding values from SNP profiles. Random forest, gradient boosting, and support vector machine models are all deployed in active chickpea breeding programs.
- Deep learning (multi-layer neural networks capable of learning complex, hierarchical patterns) is applied where the input data is high-dimensional and the relationships are non-linear, such as predicting grain yield from hyperspectral imaging data or modeling genotype-by-environment interactions across diverse growing locations.
- Computer vision (AI systems that interpret and quantify visual data from cameras and sensors) enables automated measurement of plant traits — leaf area, canopy architecture, pod number, and disease lesion severity — directly from field or greenhouse imagery, replacing slow and variable manual scoring.
- Predictive analytics platforms such as ICRISAT’s genomic selection tools integrate multi-environment trial data with genomic data to generate breeding value estimates that account for both genetic potential and environmental interaction.
A 2024 study published in Frontiers in Plant Science demonstrated a CNN (convolutional neural network) and extreme gradient boosting hybrid model that predicted seeds per plant and thousand-seed weight in chickpea with an accuracy of approximately 85%, using genomic data encoded as artificial image objects — a novel approach that treats genetic profiles as visual patterns and feeds them into image-recognition architectures (PMC, 2024).
AI-Powered Phenotyping Across Chickpea Fields
Phenotyping (the systematic measurement of observable plant characteristics) has historically been the rate-limiting step in crop breeding: growing thousands of plants and measuring them manually takes enormous time, labor, and field space. AI-powered phenotyping dissolves this bottleneck by automating measurement through remote sensing technologies.
Drone-mounted multispectral cameras capture images across visible and near-infrared wavelengths simultaneously, enabling the calculation of vegetation indices that correlate with chlorophyll content, water stress, and biomass.
When these indices are tracked across a breeding nursery over time, machine learning models can identify high-performing lines weeks before harvest — effectively giving breeders a genomics-free early warning system that complements their DNA-based predictions.
Satellite monitoring adds a landscape-level layer. Platforms like Planet Labs and ESA’s Sentinel-2 deliver repeat imagery at 3–10 meter resolution, sufficient to track canopy development across large multi-location trials.
When satellite-derived phenotypic data from multiple environments is fed into AI models alongside genomic data, the models learn how genetic potential expresses differently under different soil types, rainfall regimes, and temperature profiles — dramatically improving the accuracy of selection decisions made in target growing environments.
Mohanty, Yadav et al. (Plant, Cell and Environment, 2025) identified 10 key genomic loci linked to heat stress tolerance in chickpea using multi-locus GWAS, pinpointing hub regulatory genes including CIPK25 and WRKY40 as primary candidates for marker-assisted introgression into elite breeding lines.
Breeders in hot semi-arid regions can now screen germplasm collections at the DNA level for these heat-tolerance markers, fast-tracking the development of varieties suited to temperatures that current commercial lines cannot tolerate without yield loss.
Creating Climate-Smart Chickpeas for Specific Regions
Climate-smart breeding goes beyond simply producing stress-tolerant varieties — it involves designing cultivars specifically for the predicted future climate of a target region, not its past climate.
AI makes this possible by integrating General Circulation Model (GCM) outputs — standardized climate projections from meteorological research institutions — with crop growth models and genomic prediction platforms to define the ideal ideotype (the theoretical plant with the optimal combination of traits for a defined environment) for a given location in 2040 or 2050.
This approach is already being piloted by ICRISAT and partner national agricultural research systems (NARS) in India, Ethiopia, and Tanzania.
Breeders define a target environment profile — average maximum temperature during flowering, expected soil moisture levels during grain fill, likely pathogen pressure — and AI models search the global chickpea gene bank to identify which existing varieties or crossing combinations come closest to the predicted optimum.
This is fundamentally different from classical breeding, where varieties are adapted to current conditions and then tested retrospectively against stress events.
The sustainability dividend is equally important. Varieties optimized for specific environments require fewer inputs — less irrigation water, reduced fungicide applications for disease-resistant lines, and lower nitrogen fertilizer needs because chickpea’s root-associated nitrogen-fixing bacteria (rhizobia) can supply a significant portion of the crop’s nitrogen requirement.
AI-designed varieties that perform at high levels with minimal external inputs represent a genuine contribution to sustainable agriculture at scale.
Case Studies and Research Projects Leading the Way
The most comprehensive real-world example of AI-assisted chickpea breeding is the international pan-genome project published in Nature in 2021, involving researchers from the University of Queensland, ICRISAT, and 40 partner institutions across 13 countries.
The team sequenced 3,366 chickpea accessions, assembled a reference pan-genome capturing previously unmapped genetic diversity, and used the FastStack AI platform to model the optimal genetic architecture for maximum seed weight — a key yield proxy. Their AI-generated ideal genotype was predicted to outperform the best available commercial varieties by up to 12% for seed weight.
AI does not replace the plant breeder’s judgment — it amplifies it. The breeder defines the target, the environment, and the constraints. AI searches a solution space too vast for any human mind to navigate alone.
In Australia, the Agriculture Victoria Research program (in collaboration with La Trobe University and the University of Adelaide) applied genomic selection models to Ascochyta blight resistance breeding, demonstrating that AI-guided pre-selection could achieve prediction accuracies comparable to expensive field disease nurseries at a fraction of the cost. This finding has practical implications for any breeding program working with limited budgets and field capacity.
The University of Sydney’s Plant Breeding Institute conducted a GWAS study specifically targeting heat tolerance QTLs linked to canopy closure and early flowering, identifying markers now being actively incorporated into Australian national breeding programs.
Benefits for Farmers on the Ground
The downstream impact of AI-designed chickpea varieties reaches farmers through three primary channels: higher and more stable yields, reduced crop losses from disease and climate stress, and lower input costs from varieties that need less irrigation, fewer chemical inputs, and less fertilizer to perform at their genetic potential.
- Yield gains of 10–23% — as modeled in the Nature 2021 pan-genome study — translate directly into farm income in markets where chickpea is a cash crop, providing smallholders with a meaningful improvement in economic resilience.
- Disease-resistant varieties reduce the frequency and severity of crop failures, which in regions where chickpea is a primary household income source can mean the difference between a manageable loss and a debt spiral.
- Water-use-efficient varieties allow farmers in semi-arid regions to grow chickpea with 20–30% less irrigation water compared to conventional varieties, a critical advantage where groundwater depletion is already a pressing agricultural crisis.
AI-assisted breeding also shortens the time between a farmer’s need and a breeder’s solution. When a new disease race emerges or a weather pattern shifts, AI can reorient the breeding strategy and identify new selection targets within a single season, rather than requiring a five-year research cycle to characterize the problem and a further decade of crossing programs to address it.
Benefits for Consumers and the Global Food System
From a food system perspective, more productive, resilient chickpea varieties improve both the quantity and reliability of plant-based protein supplies. The global chickpea production volume reached 19 million tons in 2024, an increase of 8.9% over the previous year (IndexBox, 2025).
Sustaining and accelerating this trajectory requires varieties that can cope with a climate that will not resemble the one under which current varieties were developed.
Higher-protein, higher-iron chickpea varieties developed through AI-guided biofortification offer a nutritional upgrade across the entire food chain — from direct consumption in South Asian and Middle Eastern households to ingredient supply chains feeding the global chickpea pasta, hummus, and plant-protein concentrate industries.
As the global chickpea market grows toward a projected USD 29.59 billion by 2033 (SkyQuest, 2025), the nutritional quality of the underlying crop becomes a competitive differentiator in premium food markets while remaining a public health lever in subsistence farming economies.
Supply chain stability also improves. When varieties are specifically designed for a target environment and possess multiple layers of stress resilience, the year-to-year variability in production volumes decreases, smoothing price volatility for food processors, traders, and end consumers alike.
Challenges and Limitations That Must Be Acknowledged
AI-assisted chickpea breeding is genuinely transformative, but it carries real constraints that prevent it from being a universal solution available to all breeding programs immediately.
1. Data quality requirements are demanding. AI models are only as accurate as the training data they learn from. If phenotypic records are inconsistently collected, if trial management varies between locations, or if genotyping errors contaminate the SNP datasets, prediction accuracy degrades — sometimes to the point where AI-guided selections perform no better than random choices.
2. Access to AI technologies remains unequal. National agricultural research systems in low-income countries — precisely the places where resilient chickpea varieties are most urgently needed — often lack the computational infrastructure, bioinformatics capacity, and digital data management systems required to implement genomic selection programs independently.
3. Research and development costs for high-throughput genotyping and phenotyping platforms remain substantial, even as per-sample sequencing costs have fallen dramatically. Building and maintaining the reference datasets, platform software, and analytical pipelines requires sustained institutional investment that many programs cannot guarantee.
4. Regulatory and ethical considerations arise where AI-designed varieties are developed using germplasm accessed from international gene banks, triggering questions about benefit-sharing obligations under the International Treaty on Plant Genetic Resources for Food and Agriculture (ITPGRFA). Transparency in how AI selects among genetic resources and equitable distribution of the resulting varieties are active areas of policy discussion.
Jeffrey, Kaiser, Trethowan and Ziems (Frontiers in Plant Science, 2024) conducted a GWAS across diverse chickpea germplasm and identified QTLs for canopy closure and early flowering that collectively explain a significant proportion of phenotypic variance for heat avoidance traits, validating these loci as targets for marker-assisted and genomic selection in warm-climate breeding programs.
Breeders working in regions where temperatures regularly exceed 35°C during pod-fill now have validated genetic markers to accelerate the development of commercial heat-escape varieties without sacrificing yield potential.
The Future of AI-Designed Chickpeas
The next phase of AI integration in chickpea breeding is converging around several complementary developments. Digital breeding platforms — cloud-based environments where genomic data, phenotypic records, climate projections, and AI prediction models coexist and interact — are becoming the standard operating environment for advanced breeding programs.
These platforms allow breeders to test hypothetical crossing schemes in silico (computationally, without growing a single plant) before committing expensive resources to physical trials.
The breakthrough is not the AI tool itself — it is the marriage of genomic data, environmental intelligence, and predictive modeling into a single decision framework that no previous generation of breeders has had access to.
Speed breeding technologies, which use extended photoperiod and controlled temperature environments to compress the chickpea life cycle to as little as 8–10 weeks per generation (versus the 20–28 weeks of a normal field cycle), are increasingly paired with AI-guided selection to further accelerate the breeding pipeline.
Combined, these two technologies can in principle deliver a new chickpea variety from concept to commercial release in 4–6 years, compared to the 10–15 years typical of conventional programs.
Precision agriculture integration will take AI-designed varieties and match them to farm-level soil and microclimate data, enabling seed companies to recommend the optimal variety for a specific farmer’s field rather than for a broad regional average.
And as genome-editing technologies like CRISPR-Cas9 continue their regulatory journey toward commercial adoption, AI models will serve as the design layer — predicting which precise edits to the chickpea genome will deliver the target trait profile in a particular environmental context.
Conclusion
AI helps design the perfect chickpea not as a final product but as an ongoing, adaptive process — one that continuously incorporates new genetic data, new climate projections, and new agronomic insights to stay ahead of the challenges facing the crop and the farmers who grow it. The global chickpea market, now valued at USD 15.96 billion and growing at 7.1% annually, is both a commercial opportunity and a food security imperative, and the varieties that will serve it in 2035 and beyond are being designed today on computational platforms that did not exist a decade ago. The integration of AI into chickpea crop improvement is not a future possibility — it is an active, well-funded, internationally collaborative scientific enterprise already producing measurable results in fields from Australia to India to Ethiopia.
Frequently Asked Questions (FAQs)
Can AI create entirely new chickpea varieties from scratch? AI does not create new varieties independently. It analyzes existing genetic diversity, predicts which crossing combinations will produce the best performing offspring, and guides breeders toward faster and more accurate selection decisions. The biological work of crossing, growing, and evaluating plants still occurs in the physical world — AI accelerates and optimizes the decisions made within that process.
How accurate are AI breeding predictions in chickpea? Accuracy depends heavily on the size and quality of the training dataset and the genetic complexity of the trait being predicted. For relatively simple traits with well-characterized genetic architectures, current AI models achieve prediction accuracies of 0.6–0.85 (measured as correlation between predicted and observed values). For highly complex traits like overall yield in diverse environments, accuracies are typically lower and improving as datasets grow.
Are AI-designed chickpeas genetically modified? No. AI-assisted genomic selection and conventional marker-assisted breeding do not involve modifying the plant’s DNA — they involve selecting among naturally occurring genetic variations that already exist in the chickpea gene pool. The resulting varieties are bred, not engineered, and are not classified as genetically modified organisms (GMOs) under current regulatory frameworks in most countries.
How long does AI-assisted chickpea breeding take? With AI-guided genomic selection and speed breeding technologies working in combination, the development timeline for a new variety can be compressed to 4–6 years, compared to 10–15 years for conventional breeding programs. However, regulatory variety release processes, seed multiplication timelines, and national approval requirements add time beyond the breeding phase itself.
Will AI make chickpea farming more sustainable? Evidence strongly supports this outcome. Varieties with higher yield stability require fewer emergency inputs during stress years, varieties with disease resistance reduce fungicide application frequency, and drought-tolerant varieties reduce irrigation demand in water-scarce regions. All three of these outcomes move chickpea farming toward more sustainable resource use — which is precisely the direction that both environmental necessity and consumer expectations are demanding.
References:
1. Guo, T. (2024). Genetic improvement of chickpeas: traits, targets, and technology. Legume Genomics and Genetics, 15.
2. Sun, L., Lai, M., Ghouri, F., Nawaz, M. A., Ali, F., Baloch, F. S., … & Shahid, M. Q. (2024). Modern plant breeding techniques in crop improvement and genetic diversity: From molecular markers and gene editing to artificial intelligence—A critical review. Plants, 13(19), 2676.
3. Rai, K. K. (2022). Integrating speed breeding with artificial intelligence for developing climate-smart crops. Molecular biology reports, 49(12), 11385-11402.
4. Karakoy, T., Yelmen, I., Zontul, M., & Yildirim, F. (2026). Predicting Chickpea Yield Using Artificial Neural Networks with Explainable AI. Agronomy, 16(7), 768.
5. Altaf, M. T., Liaqat, W., Ali, A., Jamil, A., Fahad, M., Rahman, M. A. U., … & Mohamed, H. I. (2025). Advancing chickpea breeding: omics insights for targeted abiotic stress mitigation and genetic enhancement. Biochemical Genetics, 63(2), 1063-1115.
6. Wang, K., Xia, L., Yang, X., Du, C., Tang, T., Yang, Z., … & Zhang, J. (2025). Integrating Artificial Intelligence and Biotechnology to Enhance Cold Stress Resilience in Legumes. Plants, 14(17), 2784.
7. Roorkiwal, M., Bharadwaj, C., Barmukh, R., Dixit, G. P., Thudi, M., Gaur, P. M., … & Varshney, R. K. (2020). Integrating genomics for chickpea improvement: achievements and opportunities. Theoretical and Applied Genetics, 133(5), 1703-1720.
8. Bose, S., Banerjee, S., Kumar, S., Saha, A., Nandy, D., & Hazra, S. (2024). Review of applications of artificial intelligence (AI) methods in crop research. Journal of Applied Genetics, 65(2), 225-240.
9. Singh, S., Vij, S., & Kushwah, A. (2026). Harnessing speed breeding: a pathway to accelerating genetic gain in crop breeding. Journal of Plant Biochemistry and Biotechnology, 1-17.
10. Shahnaz, Z., Zia, M. A. B., Shahnaz, A., Manzoor, Z., Ismail, A., & Hassan, Z. (2025). Chickpea transcriptomics: insights into stress responses and future applications. Crop & Pasture Science, 76(5), CP24293.
11. Sheershika, S., Prasanna, G., Narkhede, G., Delvadiya, I. R., & Jarial, S. Genomic-assisted Breeding and Wild Germplasm Utilization for Climate-resilient Chickpea (Cicer arietinum L.): Advances and Future Perspectives: A Review. Legume Research-An International Journal, 1, 11.
12. Jha, U. C., Naik, Y. D., Priya, M., Nayyar, H., Sofi, P. A., Beena, R., … & Siddique, K. H. (2025). Chickpea (Cicer arietinum L.) battling against heat stress: plant breeding and genomics advances. Plant Molecular Biology, 115(4), 101.
