Wheat (Triticum aestivum L.), a globally cultivated staple crop, serves as a primary food source for approximately 35% of the world’s population. Given its status as the most abundant and traded food commodity, its economic significance cannot be understated (Guo et al., 2018). The spike, an essential reproductive structure of wheat, assumes a critical role in grain production and maintenance (Faris et al., 2014). Previous investigations have extensively explored the association between spike morphology and seed yield, specifically focusing on spike densities (SD), spike length (SL), and spikelet number per spike (SNS) (Chai et al., 2019; Li T. et al., 2021; You et al., 2021; Yu et al., 2022). The shape of the wheat spike exhibits a direct correlation with yield and is thus considered a pivotal agronomic trait during the process of domestication and selective breeding of wheat. Notably, the characteristics of the spike, including grain number per spike, exert a direct influence on yield levels. Therefore, gaining a comprehensive understanding of the genetic regulation underlying grain number per spike in wheat holds immense theoretical significance and practical value (Cui et al., 2012).

A comprehensive understanding of the physiological, genetic, and developmental mechanisms governing spikelet morphology is of paramount importance, as it not only contributes to the augmentation of spikelet numbers but also facilitates the improvement of spikelet fruiting or setting rates (Slafer et al., 2015; Gámez et al., 2020). However, conventional breeding techniques employed to enhance wheat yields encounter significant challenges, given the considerable costs and labor-intensive nature associated with incorporating these traits into extensive germplasm collections (Cui et al., 2012). Consequently, there arises a pressing need to explore innovative methodologies that can effectively elevate wheat yield potential.

The utilization of molecular markers plays a pivotal role in molecularly assisted selection (MAS) for plant breeding, as it establishes a link between phenotypes and molecular markers. Over time, various types of molecular markers have been developed, encompassing the first-generation markers such as restriction fragment length polymorphism (RFLP), random amplified polymorphic DNA (RAPD), and amplified fragment length polymorphism (AFLP), as well as the second-generation markers like simple sequence repeats (SSR) and inter-simple sequence repeats (ISSR). Nonetheless, these markers exhibit limitations in terms of throughput, accuracy, time consumption, labor intensiveness, and cost (Zheng et al., 2018; Clifton-brown et al., 2019; Chen et al., 2022). To surmount these challenges, a third generation of molecular markers known as SNPs has emerged. SNPs are DNA sequence variants that arise from a single nucleotide alteration in the genome sequence. They are typically distributed extensively across the genome and represent the most abundant form of genomic polymorphism, thereby making SNP markers the densest available marker type. The utilization of high-throughput DNA sequencing technologies enables the rapid and efficient identification of a vast number of SNPs in species, facilitating genotyping at high marker densities (Graner et al., 2013). Moreover, SNPs are often employed for population structure determination and linkage disequilibrium (LD) analysis (Jain et al., 2014). Concurrently, insertion/deletion (InDel) markers, which are also employed for fine mapping and marker-assisted selection in diverse crops (Liang et al., 2011; Wang et al., 2012; Das et al., 2015), can exert an influence on gene expression and function, given their location within coding and regulatory regions (Jain et al., 2014). Variants present within these regions can induce alterations in protein function and modulate gene expression, as exemplified by SNPs located on exons of the GA2ox gene, impacting spikelet sterility in wheat (Alqudah et al., 2020). Consequently, the discovery of SNPs assumes paramount significance for investigating genomic variation in crop species (Shen et al., 2017). In a genome-wide association study conducted on wheat, ten stable SNPs were identified, demonstrating associations with seeding emergence rate (SER) and tiller number at various fertility stages. Ultimately, a subset of these alleles exhibited efficacy in enhancing the seeding emergence rate and tiller number at diverse fertility stages (Chen et al., 2017).

SNPs have emerged as a highly effective tool for plant breeding through the implementation of MAS. However, the application of SNP-based array technology presents certain limitations in the detection of novel loci. To address this, the development of SLAF-seq markers has provided a promising solution, offering advantages such as deep sequencing, reduced sequencing costs, optimized marker efficiency, and suitability for large populations (Sun et al., 2013; Wang et al., 2022). In conjunction with BSA techniques, the SLAF-BSA method has been successfully employed for the identification of major QTLs associated with specific traits in various plant species. Unlike individual trait analysis, the BSA technique circumvents this requirement by focusing on individuals exhibiting contrasting extreme phenotypes. This integrated approach offers a cost-effective and accurate means to elucidate the genetic architecture underlying target traits in plant species lacking a reference genome (Michenlmore and Kesseli, 1991; Du et al., 2019). The efficacy of the SLAF-BSA method has been demonstrated across diverse plant species, including peas, oilseed rape, and pepper, where it has successfully identified QTL profiles and candidate genes associated with specific traits such as leaf shape, seed weight, first flowering node, and resistance to blast root rot. Thus, the SLAF-BSA method emerges as a potent tool for the discovery of genetic markers linked to desirable traits, ultimately facilitating plant breeding endeavors (Geng et al., 2016; Xu et al., 2016; Zhang et al., 2018; Zheng et al., 2018).

Pubing3228, a novel wheat germplasm derived from the hybridization of common wheat and Agropyron, exhibits numerous superior agronomic traits, particularly in relation to spike characteristics. However, the molecular basis governing the spike traits of Pubing3228 remains largely unexplored. Therefore, the primary objective of this investigation was to establish a high-resolution genetic map utilizing the SLAF-BSA approach in two wheat germplasms, namely Pubing3228 and Jing4839, in addition to F₂ population generated from two exceptional wheat cultivars. By analyzing the extreme spike types observed in the two wheat germplasms, we aimed to identify stable molecular markers and candidate regions with the potential to contribute to future endeavors in fine mapping and gene cloning, specifically targeting key genes associated with enhanced wheat yield. This study represents a significant advancement in our comprehension of the genetic mechanisms underlying wheat yield traits, thereby paving the way for the development of more efficient strategies for wheat breeding and cultivation.

Significant distinctions were observed between Pubing3228 and Jing4839 in terms of spike length, morphology per spikelet, awning length, and number of grains per spike, as depicted in Figures 1A, B . Pubing3228 exhibited longer spikes, larger morphology per spikelet, and an absence of awning throughout the spike, whereas Jing4839 displayed smaller spikes, smaller morphology per spikelet, and longer awning throughout the spike. The average grain count per spike was determined to be 88 for Pubing3228 and 32 for Jing4839.

In order to identify potential molecular markers and candidate regions associated with the target trait, a mixed pool sample was created by crossing the two wheat varieties, and the resulting F₂generation was classified into two distinct types based on grain number. The multi-spike grain type exhibited an average of 80 grains per spike, whereas the few-spike grain type had an average of 36 grains per spike, as illustrated in Figure 1C .The statistical information regarding the seed grain situation of the F₂ generation is presented in Supplementary Table 1 . Subsequently, the second generation progeny was subjected to sequencing and subsequent analysis to construct the mixed pool sample.

In wheat breeding programs, the identification of genes or QTL associated with yield-related traits is of great importance, as these traits often exhibit higher heritability compared to yield itself. This genetic information is crucial for understanding the underlying mechanisms of yield and facilitating the genetic improvement of wheat varieties (Li et al., 2022). Among the yield-related traits, spike-related traits hold promise for enhancing seed yield. However, conventional screening of these traits in large isolated populations can be laborious, costly, and inefficient. As a result, the adoption of molecular-assisted selection (MAS) or genetic selection has gained popularity as an efficient breeding approach. A comprehensive understanding of the genetic basis underlying yield traits, including wheat spike morphology, is essential for enhancing wheat yield. In this regard, the utilization of genetic resources such as wheat genome sequences and SNP platforms for wheat breeding plays a pivotal role in the development of improved wheat varieties (Malik et al., 2021).

SLAF-seq, a cost-effective technique, has proven to be highly valuable in generating a large number of polymorphic markers, thereby enabling the construction of high-density genetic maps for plant species with extensive genomes (Sun et al., 2013). Compared to traditional PCR-based methods, numerous studies have demonstrated that SLAF-seq significantly enhances the accuracy and efficiency of QTL mapping (Zheng et al., 2018). To improve mapping precision and narrow down candidate regions, BSA has been employed, which involves the selection of individuals with extreme phenotypes. BSA has been successfully utilized as a molecular marker in various organisms, and protocols have been developed specifically in model plants such as Arabidopsis and rice (Song et al., 2017). Classical BSA analysis reduces the cost of second-generation sequencing by pooling equimolar amounts of DNA from individuals exhibiting the same trait (Zhu et al., 2021). In our study, we performed SLAF-seq and obtained a total of 2,810,474 SLAF tags, resulting in the identification of 187,489 SNPs between the parental lines and 164,018 SNPs between the mixed pools. Using the SNP-index association algorithm, we identified 24 trait-associated marker regions spanning a cumulative length of 44.33 Mb. Additionally, employing the ED association algorithm led to the identification of 70 trait-associated regions. Within these association regions, a total of 399 genes were annotated. Notably, three genes harbored non-synonymous mutant SNP loci, suggesting a direct association with the trait under investigation. These genes were considered as promising candidates for subsequent analyses.

Further analysis of the 399 annotated genes revealed that the KEGG enrichment results highlighted the enrichment of ABC transporters, which are responsible for facilitating the transportation of various substances across membranes. Among these transporters, certain inward transporters play a vital role in the transport of nutrients, including amino acids and sugars, from the extracellular environment to the intracellular matrix. This process promotes cell growth and viability (Theodoulou and Kerr, 2015). Additionally, several amino acid metabolite pathways, such as Glycine, serine, and threonine metabolism, Alanine, aspartate, glutamate metabolism, Arginine and proline metabolism, and Arginine biosynthesis, were identified through KEGG enrichment analysis. Furthermore, specific metabolites associated with cell wall, respiratory, and protective functions were found to be correlated with genotypic superiority and yield stability (Vergara-Diaz et al., 2020).

Environmental factors, including temperature, photoperiod, water, and mineral availability, have been shown to regulate spikelet differentiation rate, apical spikelet formation, and ultimately determine the final grain number. Carbon fixation and metabolic pathways were also enriched in our analysis. Gámez et al. (2020) demonstrated that a significant portion of the carbon accumulated in wheat seeds is derived from the spikelet. They further found that the spikelet exhibits twice the capacity of the flag leaf lamina in re-fixing respired CO₂ and shows higher rates of gross photosynthesis and respiration compared to the flag leaf.

The GO analysis revealed the enrichment of terms related to hormones, kinases, and floral development, particularly in response to growth hormone (GO:0060416). This finding underscores the role of growth hormone, a key plant hormone, in promoting cell polarity establishment, cell elongation, and influencing spike type and the number of grains per spike (Gallavotti et al., 2008). Moreover, auxin transport proteins were identified as critical for auxin distribution during wheat spike development, suggesting their significance in regulating this process (Li et al., 2018).

The inflorescence branching pattern in wheat spikes is a crucial trait that significantly impacts floret formation and overall productivity. Transcription factor annotation of genes within candidate regions has revealed a noteworthy similarity to members of the APETALA2 (AP2)-like transcription factor family, particularly the gene TraesCS3A01G259900, which bears a close relationship to the Q gene (Simons et al., 2006). The Q gene is considered a major gene in wheat domestication, and its q allele is prevalent in cultivated wheat varieties. Extensive research has highlighted the q allele’s influence on various traits, including subcompact spike phenotype, rachis fragility, gum toughness, plant height, and flowering time. Moreover, it is associated with deltoid spike morphology and non-free-threshing grains (Faris et al., 2005).Loss-of-function alleles at the q locus in wheat exhibit an extended developmental stage of the rachilla, leading to a significant increase in the number of florets per spikelet. MicroRNA172 (miRNA172) interactions with its target genes, particularly the genes encoding AP2-like transcription factors such as the Q gene, play a pivotal role in cereal inflorescence development and structure (Debernardi et al., 2017; Greenwood et al., 2017). The AP2 transcription factor family, including the Q gene (AP2L5) and its associated paralog AP2L2, is regulated by miRNA172 and demonstrates critical and redundant functions in specifying axillary floral meristems (Debernardi et al., 2020). The proper balance of AP2-like gene expression, located on chromosomes 5AL (Q), 2AL (AP2L2-2A), 2BL (AP2L2-2B), and 2DL (AP2L2), is essential for spikelet and floret development (Adonina et al., 2021). Structural reorganization of a chromosome during hybridization can disrupt this balance, thereby offering an opportunity to modify spikelet structure and potentially improve grain yield by manipulating this regulatory module.TheWFZP gene, a member of the AP2/ERF transcription factor family, has been experimentally demonstrated to regulate spikelet development and exhibits conserved functionality between wheat and Brachypodium. Mutations in WFZP can affect spikelet development by activating TaGW5 or directly or indirectly inhibiting TaGW8, thereby influencing grain width and weight. Enhancing wheat grain yield remains a challenging task in breeding programs, and the identification of valuable genes and favorable alleles through screening represents a crucial approach towards achieving this goal (Li T. et al., 2021).

The original contributions presented in the study are publicly available. This data can be found here: National Genomics Data Center, accession number CRA015054.

JW: Conceptualization, Formal analysis, Funding acquisition, Investigation, Software, Supervision, Validation, Writing – original draft. EW: Data curation, Investigation, Writing – review & editing. SC: Data curation, Validation, Writing – review & editing. AM: Investigation, Validation, Visualization, Writing – review & editing.