Chickpea (Cicer arietinum L.) is one of the vital food legume crops represented by two of its major cultivar types- desi and kabuli (Kumar et al., 2011) which are thought to be domesticated along with the wild ancestor C. reticulatum at Fertile Crescent around 10000 years ago (Burger et al., 2008; Toker, 2009). Development of early-flowering/maturing stress tolerant cultivars with high seed and pod yield is the prime objective of the present genomics-assisted breeding research in chickpea (Hegde, 2010; Zaman-Allah et al., 2011). The number of days to flowering is a major seed and pod-yield component trait that highly acclimatizes with climate change, diverse environmental and a/biotic stress factors and photoperiod response along with various other growth/developmental-related traits (Aryamanesh et al., 2010; Kashiwagi et al., 2013; Daba et al., 2016). Therefore, implication of flowering time in defining productivity as well as developing stress tolerant cultivars is well-documented in chickpea. Collectively, this infers that flowering time is a complex quantitative trait and it is governed by multiple major as well as minor genes/QTLs (quantitative trait loci). A strong impact of a known major evolutionary bottleneck- vernalization- on flowering time response during chickpea domestication infers that the flowering time is a most important domestication trait selected during breeding of presently cultivated desi and kabuli accessions (Abbo et al., 2014). The genetic dissection of complex flowering time quantitative trait by identifying the functionally relevant potential genes/alleles colocalized at QTLs governing this major yield component and domestication trait is thus imperative for their broader effective practical applicability in marker-aided genetic improvement of chickpea.

Significant progress has been made to decipher the complex genetic inheritance characteristics and molecular genetic dissection of flowering time trait in chickpea (Anbessa et al., 2006; Cobos et al., 2007; Pierre et al., 2008, 2011; Aryamanesh et al., 2010; Zhang et al., 2013). This involves identification of four different major early flowering (efl) gene loci/allelic variants [efl1, efl2/ppd (photoperiod), efl3 and efl4] controlling varied flowering time trait adaptation characteristics in multiple chickpea accessions (Hegde, 2010; Gaur et al., 2014; Weller and Martínez, 2015). Additionally, this includes colocalization of various known flowering time gene homologs [like Efl1, Efl2, LFY (LEAFY) and FT (flowering time) gene families] within the low-resolution major flowering time QTL regions mapped on chickpea chromosomes (Cho et al., 2002; Anbessa et al., 2006; Lichtenzveig et al., 2006; Cobos et al., 2007, 2009; Aryamanesh et al., 2010; Hossain et al., 2010; Rehman et al., 2011; Vadez et al., 2012; Jamalabadi et al., 2013; Varshney et al., 2014).

Substantial efforts have also been made to understand the complex gene regulatory networks and transcriptional modules governing flower development in a desi chickpea accession (ICC 4958) through NGS (next-generation sequencing)-based global transcriptome sequencing strategy (Singh et al., 2013). The deployment of these differentially expressed candidate gene-derived SNPs in association mapping and their subsequent integration with GWAS (genome-wide association study), high-resolution QTL mapping, differential transcript profiling, molecular haplotyping have delineated tissue/stage (flower bud/flower)-specific differentially regulated potential candidate genes underlying major QTLs regulating flowering time at a whole genome level in chickpea (Das et al., 2015b; Upadhyaya et al., 2015). Until yet, none of these identified genes harboring major flowering time QTLs have been validated in multiple genetic backgrounds (mapping populations) and identified through map-based cloning that could be employed for marker-aided genetic enhancement of chickpea. This could be restrained due to low marker genetic polymorphism particularly between parents of multiple intra-/inter-specific mapping populations along with limited accessibility of large size mapping populations and high-density genetic linkage maps of chickpea. An alternative genome-wide approach is thus essential for quick identification and molecular mapping (fine-mapping/map-based isolation) of high-resolution flowering time QTLs/genes in order to drive genomics-led crop improvement in chickpea.

For genetic mapping of major flowering time QTLs, conventional QTL mapping approach that primarily involves genotyping of large-scale SSR (simple sequence repeat) and SNP (single nucleotide polymorphism) markers among mapping individuals of diverse inter-/intra-specific populations is found much expedient in chickpea (Anbessa et al., 2006; Lichtenzveig et al., 2006; Cobos et al., 2007, 2009; Aryamanesh et al., 2010; Hossain et al., 2010; Gowda et al., 2011; Rehman et al., 2011; Vadez et al., 2012; Jamalabadi et al., 2013; Stephens et al., 2014; Varshney et al., 2014). This approach essentially identified a diverse array of low-resolution longer marker confidence interval spanning major QTLs associated with chickpea flowering and maturation time (Cho et al., 2002; Anbessa et al., 2006; Lichtenzveig et al., 2006; Cobos et al., 2007, 2009; Aryamanesh et al., 2010; Hossain et al., 2010; Rehman et al., 2011; Vadez et al., 2012; Jamalabadi et al., 2013; Varshney et al., 2014).

The freely accessible draft genome sequences are found much proficient to accelerate genome and transcriptome resequencing of diverse desi, kabuli and wild accessions that are most commonly utilized as parents for generating diverse intra- and inter-specific mapping populations of chickpea (Jain et al., 2013; Varshney et al., 2013; Parween et al., 2015). Aside genomic resources, multiple genetic resources including advanced generation recombinant inbred lines (RILs) and back-cross mapping populations as well as core/mini-core germplasm accessions exhibiting a broader range of phenotypic variation for flowering time trait are now available in chickpea (Upadhyaya et al., 2001, 2008; Gaur et al., 2014). All these available genetic and genomic resources essentially have assisted in utilization of a high-throughput NGS-based QTL-seq strategy vis-à-vis a commonly adopted traditional QTL mapping approach for fast genome-wide scanning and genetic mapping of major QTLs controlling various quantitative agronomic traits (for instance, 100-seed weight, pod number and root/total plant dry weight ratio) in chickpea (Das et al., 2015a, 2016; Singh et al., 2016).

To complement this, a multiple QTL-seq assay that relies on QTL-seq analysis in multiple mapping populations generated by inter-crossing of common parental accessions, has been developed currently as a most promising genome-wide strategy for QTL mapping at a high-resolution scale (Das et al., 2016). Essentially, multiple QTL-seq involves whole genome NGS resequencing of DNA bulks (exhibiting two utmost contrasting phenotypic traits) constituted from homozygous individuals of multiple mapping populations comprising at least single common parent. This approach is found most promising based on its potential to validate QTL-seq-derived major QTLs identified from individual preliminary as well as advanced generation intra-/inter-specific mapping population in multiple mapping populations of diverse genetic backgrounds. Moreover, utility of this approach is clearly evident from its efficacy to narrow-down each QTL-seq originated sizeable long QTL genomic intervals into functionally relevant potential candidate genes governing important agronomic traits (for instance, pod number) in chickpea (Das et al., 2016).

Considering usefulness and broader practical applicability, multiple QTL-seq assay can be employed for rapid genome-wide scanning and fine-mapping (positional cloning) of trait-linked major genes and natural allelic-variants colocalized at robust QTLs (well-validated in multiple mapping populations) in chickpea with minimal resource expenses. This will collectively enrich our understanding on complex genetic architecture and evolutionary pattern influencing flowering time quantitative trait variation during domestication of chickpea in order to expedite its genomics-assisted crop improvement. In view of afore-mentioned possibilities, a multiple QTL-seq strategy was employed in two inter- and intra-specific RIL (recombinant inbred lines) mapping populations- (C. arietinum desi accession ICC 4958 × C. reticulatum wild accession ICC 17163) and (ICC 4958 × C. arietinum kabuli accession ICC 8261)- at a genome-wide scale to delineate major genomic (gene) regions and novel natural allelic variants underlying the QTLs associated with flowering time in chickpea.

A broader phenotypic variation coupled with bi-directional transgressive segregation (normal frequency distribution) of DTF trait among RIL individuals and parents of inter (ICC 4958 × ICC 17163)- and intra (ICC 4958 × ICC 8261)-specific mapping population phenotyped in field and green house (long- and short-day) conditions at two different geographical locations/years was evident. This infers the complex genetic inheritance pattern of flowering time quantitative trait in chickpea. Therefore, genetic dissection of this complex quantitative trait employing various genomics-assisted breeding strategies is essential for genetic enhancement and to develop early flowering high seed and pod-yielding stress tolerant cultivars of chickpea during present scenario of climate change. To accomplish these, our study selectively employed a rapid, cost-efficient and NGS-led high-throughput multiple QTL-seq assay in two inter- and intra-specific RIL mapping population exhibiting a much wider flowering time trait variation including a higher heritability (consistent phenotypic expression) for flowering time in field and green house (long- and short-day) across two diverse geographical locations/years in order to identify major flowering time QTLs in chickpea.

The QTL-seq analysis in an inter-specific RIL mapping population (ICC 4958 × ICC 17163) detected 1.99 Mb and 757.7 kb two major genomic regions underlying Caq^aDTF4.1 and Caq^aDTF4.2 QTLs, respectively, mapped on chromosome 4 governing flowering time in chickpea. Like-wise, 907.1 kb and 1.39 Mb, two major genomic intervals of Caq^bDTF4.1 and Caq^bDTF4.2 QTLs, respectively, mapped on chromosome 4 were detected by QTL-seq analysis in an intra-specific RIL mapping population (ICC 4958 × ICC 8261). These analyses altogether led to identify two consensus major short physical genomic regions of 757.7 kb and 907.1 kb harboring Caq^aDTF4.2 and Caq^bDTF4.1 QTLs, respectively, on chromosome 4 of chickpea. The validation of these major DTF QTLs across two diverse intra-/inter-specific chickpea mapping population was apparent implicating the robustness of identified QTLs in regulating flowering time in chickpea. The aforesaid outcomes also inferred the efficacy of multiple QTL-seq assay to narrow-down the longer major DTF QTL intervals detected by QTL-seq in an individual mapping population into shorter major QTL regions in multiple mapping populations of chickpea. This suggests the potential utility of multiple QTL-seq over NGS-based QTL-seq assay and other conventional QTL mapping approaches in high-resolution molecular/fine mapping of major genomic regions harboring QTLs governing diverse agronomic traits including flowering time in chickpea. As per congruent physical positions (bp) on desi chromosome 4, two short interval DTF QTLs (Caq^aDTF4.2 and Caq^bDTF4.1) revealed correspondence with the two earlier identified known major flowering time QTLs (CaqDF4.1 and CaqDF4.2) that are identified and mapped on an intra-specific high-density genetic linkage map (ICC 16374 × ICC 762) of chickpea (Upadhyaya et al., 2015).

The comprehensive multiple QTL-seq analysis at Caq^aDTF4.2 and Caq^bDTF4.1 QTL regions detected novel natural allelic variants-containing two strong flowering time-associated efl1 and GI genes with highest Δ (SNP-index) of 1.0 and thereby, considered as the potential candidates for flowering time regulation in chickpea. Two potential candidate genes, efl1 and GI underlying these major QTLs (Caq^aDTF4.2/CaqDF4.1 and Caq^bDTF4.1/CaqDF4.2 detected in our present and past studies, respectively) regulating flowering time have been delineated by deploying an integrated genomics-assisted breeding strategy involving candidate gene-based trait association mapping, GWAS, QTL mapping, differential transcript expression profiling and gene-specific molecular haplotyping in chickpea (Upadhyaya et al., 2015). The potential of these identified known flowering development pathway and FT gene homologs like efl1 and GI colocalized at the major QTLs in regulating flowering time have been documented by different traditional QTL mapping studies involving diverse intra- and inter-specific mapping populations of chickpea (Cho et al., 2002; Anbessa et al., 2006; Lichtenzveig et al., 2006; Cobos et al., 2007, 2009; Radhika et al., 2007; Aryamanesh et al., 2010; Hossain et al., 2010; Gowda et al., 2011; Rehman et al., 2011; Hiremath et al., 2012; Vadez et al., 2012; Jamalabadi et al., 2013; Zhang et al., 2013; Varshney et al., 2014). Notably, functional validation and comprehensive molecular characterization of photoperiod-independent efl1 gene and photoperiod-dependent circadian-clock-related GI gene have implicated their potential involvement in regulating flowering time of legumes and Arabidopsis (Hecht et al., 2007; Liu et al., 2008; Watanabe et al., 2009, 2011, 2012; Weller et al., 2009; Kang et al., 2010; Laurie et al., 2011; Andres and Coupland, 2012; Kim et al., 2012; Pin and Nilsson, 2012; Song et al., 2013; Yamashino et al., 2013; Zhai et al., 2014; Weller and Martínez, 2015).

Despite of identifying similar flowering time efl1 and GI genes between past and present studies, we were able to discover diverse novel flowering time-regulating non-synonymous and regulatory natural SNP allelic variants (unlike our previous study by Upadhyaya et al., 2015) from the two target efl1 and GI genes that are localized in the two multiple QTL-seq derived major DTF QTL regions (Caq^aDTF4.2 and Caq^bDTF4.1) of chickpea. The detection of altogether different natural allelic variants from two similar efl1 and GI flowering genes localized at two major DTF QTL regions between past and present studies collectively infers the population/cultivar-specific genetic inheritance pattern of complex flowering time quantitative trait in diverse genetic backgrounds of chickpea. The above clues collectively suggest the accuracy, robustness and wider practical applicability of multiple QTL-seq approach for fast genome-wide scanning and mapping of high-resolution major flowering time QTLs as well as delineation of candidate genes and novel natural alleles underlying these major QTLs governing flowering time in chickpea.

The quantitative flowering time trait is primarily governed by complex regulatory networks/pathways involving a diverse array of genes in plant species including legumes (Andres and Coupland, 2012; Song et al., 2013; Weller and Martínez, 2015). The molecular haplotyping of efl1 and GI genes (detected by multiple QTL-seq) among diverse desi and kabuli cultivated and wild accessions has detected multiple novel natural allelic variants including haplotypes in these flowering time genes exhibiting varied potential characteristics for flowering time trait regulation and evolutionary pattern in domesticated chickpea (Upadhyaya et al., 2015). Therefore, novel functionally relevant potential molecular signatures (SNP markers, genes, QTLs and natural allelic variants) governing flowering time delineated by us employing a NGS-based high-throughput multiple QTL-seq strategy can be useful for fast genetic dissection of complex flowering time quantitative trait and eventually genomics-assisted crop improvement to develop early flowering varieties of chickpea with limited resource expenses.

Preliminary efforts have been made to understand the natural/functional allelic diversity-based domestication pattern of two flowering (efl1 and GI) genes localized at two major DTF QTL regions among Cicer cultivated (desi and kabuli) and wild genepools. This exhibited a significant reduction of natural/functional allelic diversity in cultivated desi and kabuli accessions from diverse geographical regions of the world as compared to annual and perennial wild accessions belonging to primary, secondary and tertiary gene pools of chickpea. However, this observed background allelic diversity reduction in cultivated than that of wild chickpea was much stronger especially at major DTF QTL regions where efl1 and GI flowering genes were mapped. This implicates that these genes were targeted by artificial selection which was further evident from their non-neutral evolution during chickpea domestication based on significant variation of Tajima’s D between cultivated (–1.12) and wild (0.26) accessions, respectively. Interestingly, all these natural allelic variants-containing two potential genes localized within flowering time major QTL intervals have been commonly identified and mapped on multiple independent chickpea mapping populations by earlier and our present studies (Upadhyaya et al., 2015). These outcomes clearly reflect the extensive contribution of four sequential evolutionary bottlenecks including vernalization and strong artificial and/or natural selection pressure on these flowering time-associated natural allelic variants of two gene loci (efl1 and GI) during chickpea domestication leading toward reduction of genetic diversity in cultivated chickpea as compared to that of wild Cicer genepool (Lev-Yadun et al., 2000; Abbo et al., 2003; Berger et al., 2005; Burger et al., 2008; Toker, 2009; Meyer et al., 2012; Jain et al., 2013; Kujur et al., 2013; Varshney et al., 2013; Saxena et al., 2014a,b).

The most crucial evolutionary bottleneck, vernalization, is a vital key module of flowering time during chickpea domestication culminating into existence of currently cultivated vernalization insensitive desi and kabuli cultivars specifically from the vernalization sensitive wild ancestor C. reticulatum (Abbo et al., 2003, 2014; Berger et al., 2005; Burger et al., 2008; Toker, 2009). The major domestication bottlenecks integrated with artificial selection including modern breeding efforts have been constantly practiced during the chickpea genetic improvement program for developing its early flowering cultivars of high seed and pod yield. These findings collectively infer that the natural allelic variation-based functional diversity scanned in the genes might be associated with flowering time trait evolution with regard to differential domestication-led bottlenecks including vernalization response in desi and kabuli cultivated and wild chickpea during domestication. Henceforth, flowering time represents a vital component of domestication trait selected during breeding and genetics of chickpea. Moreover, the major impact of long- and short-photoperiods which are the major environmental cues for determining the flowering time including time of flower initiation and/or first flower appearance especially in photoperiod-sensitive as compared to photoperiod-insensitive chickpea accessions is well documented (Daba et al., 2016). In the present study, a significant interactions between long- and short-day photoperiods and DTF trait variation observed in individuals of two RIL mapping populations across 2 years was apparent. In spite of this concern, we were able to identify functionally relevant non-synonymous/synonymous coding and regulatory SNP allelic variants from two flowering genes (efl1 and GI) localized at two major DTF QTL regions by using the long- and short-day phenotyping data of both RIL populations separately in multiple QTL-seq assay. This further infers the efficacy of strategy (multiple QTL-seq) implemented in our study to detect potential molecular signatures regulating flowering time in chickpea. It is, therefore, essential to perform a comprehensive analysis using all natural/functional allelic variants discovered and potential locus targeted by natural and/or artificial selection in two flowering genes (efl1 and GI) to delve deeper into the complex flowering time trait evolution and domestication in chickpea. This will be useful to understand the molecular mechanism influencing fixation of such complex flowering time quantitative trait in domesticated cultivars that are adapted to multiple agro-ecological regions of the world and further pave the way for genetic enhancement to develop early flowering high seed/pod-yielding varieties of chickpea amidst current climate change scenario.

RS, AD, and UB conducted all experiments and drafted the manuscript. HU, RK, PS, ST, and CB helped in development, advancement and phenotyping of mapping populations. AT and SP conceived and designed the study, guided data analysis and interpretation, participated in drafting and correcting the manuscript critically and gave the final approval of the version to be published. All authors have read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.