Key messageAnalysis of phenotypic data for 20 drought tolerance traits in 1–7 seasons at 1–5 locations together with genetic mapping data for two mapping populations provided 9 QTL clusters of which one present on CaLG04 has a high potential to enhance drought tolerance in chickpea improvement.AbstractChickpea (Cicer arietinum L.) is the second most important grain legume cultivated by resource poor farmers in the arid and semi-arid regions of the world. Drought is one of the major constraints leading up to 50 % production losses in chickpea. In order to dissect the complex nature of drought tolerance and to use genomics tools for enhancing yield of chickpea under drought conditions, two mapping populations—ICCRIL03 (ICC 4958 × ICC 1882) and ICCRIL04 (ICC 283 × ICC 8261) segregating for drought tolerance-related root traits were phenotyped for a total of 20 drought component traits in 1–7 seasons at 1–5 locations in India. Individual genetic maps comprising 241 loci and 168 loci for ICCRIL03 and ICCRIL04, respectively, and a consensus genetic map comprising 352 loci were constructed (http://cmap.icrisat.ac.in/cmap/sm/cp/varshney/). Analysis of extensive genotypic and precise phenotypic data revealed 45 robust main-effect QTLs (M-QTLs) explaining up to 58.20 % phenotypic variation and 973 epistatic QTLs (E-QTLs) explaining up to 92.19 % phenotypic variation for several target traits. Nine QTL clusters containing QTLs for several drought tolerance traits have been identified that can be targeted for molecular breeding. Among these clusters, one cluster harboring 48 % robust M-QTLs for 12 traits and explaining about 58.20 % phenotypic variation present on CaLG04 has been referred as “QTL-hotspot”. This genomic region contains seven SSR markers (ICCM0249, NCPGR127, TAA170, NCPGR21, TR11, GA24 and STMS11). Introgression of this region into elite cultivars is expected to enhance drought tolerance in chickpea.Electronic supplementary materialThe online version of this article (doi:10.1007/s00122-013-2230-6) contains supplementary material, which is available to authorized users.

Advances in next-generation sequencing and genotyping technologies have enabled generation of large-scale genomic resources such as molecular markers, transcript reads and BAC-end sequences (BESs) in chickpea, pigeonpea and groundnut, three major legume crops of the semi-arid tropics. Comprehensive transcriptome assemblies and genome sequences have either been developed or underway in these crops. Based on these resources, dense genetic maps, QTL maps as well as physical maps for these legume species have also been developed. As a result, these crops have graduated from 'orphan' or 'less-studied' crops to 'genomic resources rich' crops. This article summarizes the above-mentioned advances in genomics and genomics-assisted breeding applications in the form of marker-assisted selection (MAS) for hybrid purity assessment in pigeonpea; marker-assisted backcrossing (MABC) for introgressing QTL region for drought-tolerance related traits, Fusarium wilt (FW) resistance and Ascochyta blight (AB) resistance in chickpea; late leaf spot (LLS), leaf rust and nematode resistance in groundnut. We critically present the case of use of other modern breeding approaches like marker-assisted recurrent selection (MARS) and genomic selection (GS) to utilize the full potential of genomics-assisted breeding for developing superior cultivars with enhanced tolerance to various environmental stresses. In addition, this article recommends the use of advanced-backcross (AB-backcross) breeding and development of specialized populations such as multi-parents advanced generation intercross (MAGIC) for creating new variations that will help in developing superior lines with broadened genetic base. In summary, we propose the use of integrated genomics and breeding approach in these legume crops to enhance crop productivity in marginal environments ensuring food security in developing countries.

We developed 1108 transcription factor gene-derived microsatellite (TFGMS) and 161 transcription factor functional domain-associated microsatellite (TFFDMS) markers from 707 TFs of chickpea. The robust amplification efficiency (96.5%) and high intra-specific polymorphic potential (34%) detected by markers suggest their immense utilities in efficient large-scale genotyping applications, including construction of both physical and functional transcript maps and understanding population structure. Candidate gene-based association analysis revealed strong genetic association of TFFDMS markers with three major seed and pod traits. Further, TFGMS markers in the 5′ untranslated regions of TF genes showing differential expression during seed development had higher trait association potential. The significance of TFFDMS markers was demonstrated by correlating their allelic variation with amino acid sequence expansion/contraction in the functional domain and alteration of secondary protein structure encoded by genes. The seed weight-associated markers were validated through traditional bi-parental genetic mapping. The determination of gene-specific linkage disequilibrium (LD) patterns in desi and kabuli based on single nucleotide polymorphism-microsatellite marker haplotypes revealed extended LD decay, enhanced LD resolution and trait association potential of genes. The evolutionary history of a strong seed-size/weight-associated TF based on natural variation and haplotype sharing among desi, kabuli and wild unravelled useful information having implication for seed-size trait evolution during chickpea domestication.

We report a map of 4.97 million single-nucleotide polymorphisms of the chickpea from whole-genome resequencing of 429 lines sampled from 45 countries. We identified 122 candidate regions with 204 genes under selection during chickpea breeding. Our data suggest the Eastern Mediterranean as the primary center of origin and migration route of chickpea from the Mediterranean/Fertile Crescent to Central Asia, and probably in parallel from Central Asia to East Africa (Ethiopia) and South Asia (India). Genome-wide association studies identified 262 markers and several candidate genes for 13 traits. Our study establishes a foundation for large-scale characterization of germplasm and population genomics, and a resource for trait dissection, accelerating genetic gains in future chickpea breeding.

Deep learning models stand for a new learning paradigm in artificial intelligence (AI) and machine learning. Recent breakthrough results in image analysis and speech recognition have generated a massive interest in this field because also applications in many other domains providing big data seem possible. On a downside, the mathematical and computational methodology underlying deep learning models is very challenging, especially for interdisciplinary scientists. For this reason, we present in this paper an introductory review of deep learning approaches including Deep Feedforward Neural Networks (D-FFNN), Convolutional Neural Networks (CNNs), Deep Belief Networks (DBNs), Autoencoders (AEs), and Long Short-Term Memory (LSTM) networks. These models form the major core architectures of deep learning models currently used and should belong in any data scientist's toolbox. Importantly, those core architectural building blocks can be composed flexibly-in an almost Lego-like manner-to build new application-specific network architectures. Hence, a basic understanding of these network architectures is important to be prepared for future developments in AI.

The genome-wide discovery and high-throughput genotyping of SNPs in chickpea natural germplasm lines is indispensable to extrapolate their natural allelic diversity, domestication, and linkage disequilibrium (LD) patterns leading to the genetic enhancement of this vital legume crop. We discovered 44,844 high-quality SNPs by sequencing of 93 diverse cultivated desi, kabuli, and wild chickpea accessions using reference genome- and de novo-based GBS (genotyping-by-sequencing) assays that were physically mapped across eight chromosomes of desi and kabuli. Of these, 22,542 SNPs were structurally annotated in different coding and non-coding sequence components of genes. Genes with 3296 non-synonymous and 269 regulatory SNPs could functionally differentiate accessions based on their contrasting agronomic traits. A high experimental validation success rate (92%) and reproducibility (100%) along with strong sensitivity (93–96%) and specificity (99%) of GBS-based SNPs was observed. This infers the robustness of GBS as a high-throughput assay for rapid large-scale mining and genotyping of genome-wide SNPs in chickpea with sub-optimal use of resources. With 23,798 genome-wide SNPs, a relatively high intra-specific polymorphic potential (49.5%) and broader molecular diversity (13–89%)/functional allelic diversity (18–77%) was apparent among 93 chickpea accessions, suggesting their tremendous applicability in rapid selection of desirable diverse accessions/inter-specific hybrids in chickpea crossbred varietal improvement program. The genome-wide SNPs revealed complex admixed domestication pattern, extensive LD estimates (0.54–0.68) and extended LD decay (400–500 kb) in a structured population inclusive of 93 accessions. These findings reflect the utility of our identified SNPs for subsequent genome-wide association study (GWAS) and selective sweep-based domestication trait dissection analysis to identify potential genomic loci (gene-associated targets) specifically regulating important complex quantitative agronomic traits in chickpea. The numerous informative genome-wide SNPs, natural allelic diversity-led domestication pattern, and LD-based information generated in our study have got multidimensional applicability with respect to chickpea genomics-assisted breeding.

Chickpea is the third most important pulse crop worldwide. Changes in cropping system that necessitate late planting, scope for expansion in rice fallows and the global warming are pushing chickpeas to relatively warmer growing environment. Such changes demand identification of varieties resilient to warmer temperature. Therefore, the reference collection of chickpea germplasm, defined based on molecular characterization of global composite collection, was screened for high temperature tolerance at two locations in India (Patancheru and Kanpur) by delayed sowing and synchronizing the reproductive phase of the crop with the occurrence of higher temperatures ($ 358C). A heat tolerance index (HTI) was calculated using a multiple regression approach where grain yield under heat stress is considered as a function of yield potential and time to 50% flowering. There were large and significant variations for HTI, phenology, yield and yield components at both the locations. There were highly significant genotypic effects and equally significant G £ E interactions for all the traits studied. A cluster analysis of the HTI of the two locations yielded five cluster groups as stable tolerant (n ¼ 18), tolerant only at Patancheru (n ¼ 34), tolerant only at Kanpur (n ¼ 23), moderately tolerant (n ¼ 120) and stable sensitive (n ¼ 82). The pod number per plant and the harvest index explained $60% of the variation in seed yield and $ 49% of HTI at Kanpur and $80% of the seed yield and $35% of HTI at Patancheru, indicating that partitioning as a consequence of poor pod set is the most affected trait under heat stress. A large number of heat-tolerant genotypes also happened to be drought tolerant.

We identified 44844 high-quality SNPs by sequencing 92 diverse chickpea accessions belonging to a seed and pod trait-specific association panel using reference genome- and de novo-based GBS (genotyping-by-sequencing) assays. A GWAS (genome-wide association study) in an association panel of 211, including the 92 sequenced accessions, identified 22 major genomic loci showing significant association (explaining 23–47% phenotypic variation) with pod and seed number/plant and 100-seed weight. Eighteen trait-regulatory major genomic loci underlying 13 robust QTLs were validated and mapped on an intra-specific genetic linkage map by QTL mapping. A combinatorial approach of GWAS, QTL mapping and gene haplotype-specific LD mapping and transcript profiling uncovered one superior haplotype and favourable natural allelic variants in the upstream regulatory region of a CesA-type cellulose synthase (Ca_Kabuli_CesA3) gene regulating high pod and seed number/plant (explaining 47% phenotypic variation) in chickpea. The up-regulation of this superior gene haplotype correlated with increased transcript expression of Ca_Kabuli_CesA3 gene in the pollen and pod of high pod/seed number accession, resulting in higher cellulose accumulation for normal pollen and pollen tube growth. A rapid combinatorial genome-wide SNP genotyping-based approach has potential to dissect complex quantitative agronomic traits and delineate trait-regulatory genomic loci (candidate genes) for genetic enhancement in crop plants, including chickpea.