Crop wild relatives (CWRs) have been established as valuable sources for incorporating genes for adaptation in crop plants under changing climatic conditions by allowing modifications in crop genetic constitution and developing high-yielding crop varieties (Hajjar and Hodgkin, 2007; Bohra et al., 2022). Among CWRs, Brassica wild relatives (BWRs) have maintained valuable genetic diversity for improving yield-related and climate-resilient traits, such as plant vigor, architecture, and the ability to withstand biotic and abiotic stresses. By enriching the gene pool of crop species, BWRs have contributed significantly by augmenting the genetic diversity of cultivated plants (Quezada-Martinez et al., 2021). The family Brassicaceae includes several wild genera, such as Brassica fruticulosa, Camelina sativa, Diplotaxis catholica, D. erucoides, Erucastrum abyssinicum, Erucastrum cardaminoides, Eruca sativa, Moricandia arvensis, and Sinapis alba. Among these, S. alba (white mustard), commonly grown as a condiment, has the highest number of PB and SB (Kumari et al., 2011; Singh et al., 2023). Interestingly, S. alba is diploid (2n= 24, SS) and has close genetic proximity to Brassica nigra L. (BB, 2n=16), a diploid ancestor of the Indian mustard (Warwick and Black, 1991).

Indian mustard (B. juncea; AABB; 2n=36) is an important oilseed crop cultivated globally for its edible oil, vegetables, and condiments (Banga and Banga, 2016). The genetic diversity of Brassica oilseed crops has become limited due to strong selection pressure for yield and quality attributes, which has made it increasingly challenging to further increase yields. Consequently, the production levels of these crops have plateaued (Singh et al., 2021a). Therefore, the enhancement of crop yield depends on the introduction of yield-related alien genes to augment the genetic basis. However, the introgression of alien genes is a tedious and time-consuming process (Kumari et al., 2020b, 2020c). Thus, the closest genera of the crop are a more reliable source for gene introduction due to the inhibition of the introduction of alien genes by intergeneric incompatibility, hybrid sterility, and reduced or absent chromosome pairing between alien and crop species and the resulting linkage drag (Bohra et al., 2022). Subsequent generations of hybrids and early backcross progenies exhibited high levels of male and female sterility due to abnormal meiosis (Lelivelt et al., 1993). In vitro fusion or protoplast fusion was initially used to introduce genes to overcome the pre- and postfertilization barriers within intergeneric hybridizations that contributed to biotic and abiotic stress tolerance in Indian mustard (Kirti et al., 1992, 1995). Moreover, genetic engineering-induced and spontaneous mutations have also been used to introduce genetic diversity. However, due to the amphidiploid nature of B. juncea, the application of reverse genetics to identify mutations that cause quantifiable phenotypic impacts is challenging. Since each phenotype is controlled by at least two mustard homoeologs with presumably redundant activities, combining mutant homoeologs from both subgenomes is necessary to change a monogenic trait (Wells et al., 2014; Emrani et al., 2015). For the trait regulated by an oligogenic or polygenic system in polyploid species, genome editing using CRISPR-Cas9 is not a realistic approach (Endo et al., 2015).

To overcome the consequences of intergeneric hybridization, we used PEG-mediated protoplast fusion to introduce new alleles of S. alba into B. juncea to avoid potential negative effects such as intergeneric incompatibility and hybrid sterility. Fortunately, we succeeded in achieving the first stable and fertile somatic hybrids of S. alba and B. juncea with proper meiosis (Kumari et al., 2018; Kumari and Bhat, 2019, 2021). These hybrids were then backcrossed with B. juncea as the recurrent parent, resulting in the first backcross progeny with a haploid set of S. alba chromosomes with a high degree of gamete viability and subsequent fertility. Among the BC₁F₂ generation, we identified a backcross line with the highest number of PBs (26–27) and SB (166) (Kumari et al., 2020b). The shoot branching pattern is controlled at various levels, such as node pattern, meristem determination, and axillary meristem elongation (McSteen and Leyser, 2005). Several genes regulate node patterns in A. thaliana, such as LATERAL SUPPRESSOR (LAS) (Greb et al., 2003), SHOOT MERISTEMLESS (STM) (Long et al., 1996), REVOLUTA (REV) (Talbert et al., 1995), and the REGULATORS OF AXILLARY MERISTEMS (RAX) genes (Keller et al., 2006; Muller et al., 2006). The inflorescence meristem identity is determined by the floral identity genes TERMINAL FLOWER1 (TFL1) (Bradley et al., 1997) and LEAFY (LFY) (Weigel et al., 1992). Moreover, branch elongation is regulated by several phytohormones, such as auxin, cytokinin, and abscisic acid (Ward and Leyser, 2004). Critical analyses of genes that regulate auxin signaling and transport, such as AUXIN RESISTANT1 (AXR1) (Lincoln et al., 1990; Leyser et al., 1993; Stirnberg et al., 1999) and MORE AXILLARY GROWTH (MAX) (Stirnberg et al., 2002; Sorefan et al., 2003; Booker et al., 2005; Bennett et al., 2006), have shown that these hormones play a central role in branch development.

However, a major challenge was the identification of the genes responsible for the segmental introgressions of B. juncea chromosomes, which were introduced in advance BCILs. To address this issue, we developed a set of SSR markers specific to the S. alba genome from the draft assembly and used them to genotype core sets of BCILs (Kumari et al., 2020a; Singh et al., 2021b, 2022b). Using a set of 170 monoallelic SSR markers specific to S. alba, we identified associated markers with high PB and SB by using genotypic and phenotypic data from two successive years. We utilized association mapping, which is a reliable method for addressing quantitative variations, to better understand introgressed variants (Risch and Merikangas, 1996; Nordborg and Tavaré, 2002; Gupta et al., 2014). The BCILs used in this study have euploid chromosomal counts high fertility, and are stabilized as translocation homozygotes.

Although the size of the introgressed segment likely plays a crucial role in defining BCILs, physical linkage significantly impacts the linkage disequilibrium (LD) between molecular markers (Remington et al., 2001). This provided a genetic basis for the association mapping of genes responsible for a substantial proportion of PB and SB. The extent of LD between linked markers in the complete set of BCILs was notably greater than that between unlinked markers. This study reports the results of association mapping experiments using markers specific to the donor genome and phenotypic data that were subsequently recorded from BC₂F_6-7 generations from two successive crop seasons. We identified closely linked markers to pinpoint the genetic regions responsible for PB and SB. The discovery of these markers will facilitate the rapid introduction of the higher branching phenotype into new germplasms, marker-assisted breeding, and selection.

During the CS-I and CS-II crop seasons at the ICAR-Directorate of Rapeseed Mustard Research, phenotypic data were collected from backcross introgression lines (BCILs) of S. alba and B. juncea. The PB and SB numbers were recorded and showed considerable variability among the BCILs. For example, BCIL 70 had an average minimum of 2.0 PB, while the donor parent S. alba in CS-II had an average maximum of 21.0 PB. The grand mean values for the number of PB were 6.32 ± 0.77 and 6.48 ± 0.45 for CS-I and CS-II, respectively. The donor parent S. alba had the highest average number of SB at 43.6, followed by BCILs 10 (26.2) and 83 (24), while BCILs 49 (5), 32 (6.6), 28 (7), 66, and 70 (7.2) had the lowest average number of SB. The grand mean values for the number of SB were 14.314 ± 3.06 and 14.36 ± 1.42 for CS-I and CS-II, respectively ( Figures 1A, B ). The cumulative effect of genotypic and environmental variance for both crop seasons resulted in the phenotypic variance of PB and SB, respectively. The phenotypic coefficient of variation (PCV) was greater than the genotypic coefficient of variation (GCV) for both traits, indicating a significant environmental influence. However, both PB and SB exhibited high broad-sense heritability (H²) values, suggesting that these traits are not influenced by environmental factors. The genetic variability for these traits was low, with genetic advancement percentages of the mean (GAM) determined to be 36.67 and 51.30 for PB and 36.55 and 69.31 for SB, respectively ( Table 1 ). The correlation between PB and SB in both seasons was very high, as shown in the correlation heatmap ( Figure 2 ).

Plant architecture plays a crucial role in determining yield potential, and thus, a positive correlation between high yields and a greater number of PB and SB has been established in Indian mustard (Ehrenreich et al., 2007). Traits that have a large genetic component and a direct correlation with yield, such as the number of branches and pods per plant, number of seeds per pod, seed weight, and flowering time, can be used as suitable selection criteria. Molecular mapping has been used to analyze these morphological characteristics to identify the genes that control them. The yield potential of B. juncea has reached a plateau due to changing climatic conditions and genetic uniformity, which are likely to worsen due to various biotic and abiotic stresses, such as insects, diseases, drought, salinity, and high temperature. The expansion of agricultural land is not a feasible solution; hence, increasing crop production through higher yields is essential. The Indian mustard has a narrow genetic base, which makes it difficult to improve its yield potential through interspecific hybridization. Therefore, it is essential to investigate genetic or allelic variability to significantly improve the yield of B. juncea. Therefore, alien gene introgression is required to strengthen the genetic base of the crop Brassicas. However, it is challenging to introgress alien genes into crop species due to various barriers, including pre- and postfertilization incompatibility, poor or lack of pairing between chromosomes, hybrid sterility, and consecutive linkage drag (Kirti et al., 1992; Bohra et al., 2022). S. alba has been used as a resistance source for biotic and abiotic stresses and seed color for almost four decades (Primard et al., 1988; Hansen and Earle, 1997; Wang et al., 2005; Kumari and Singh, 2019; Kumari et al., 2020c, 2023; Singh et al., 2021a, 2022b). To date, the S. alba genome has not been utilized for the introgression of genes related to plant architecture. The identification of alien genes that govern branching phenotypes is also difficult because the introgression of alien genes has still not been reported. However, these genes were identified in the model plant A. thaliana, and it has been confirmed that these traits are controlled by multiple genes (Ehrenreich et al., 2007).

Notably, the introgression of alien genes into a crop plant is an intricate and slow process that requires a significant amount of time and effort. To address the issue of poor or lack of pairing between chromosomes of the two species, we used S. alba as a donor for genes that govern yield-contributing traits due to its potential for greater branching (Kumari et al., 2011, 2020b). S. alba and B. nigra (a diploid progenitor of B. juncea) share the common ‘Nigra’ lineage of the subtribe Brassicinae (Warwick and Black, 1991; Nelson and Lydiate, 2006), which makes them more suitable for use as protoplast fusion partners of B. juncea. In an experiment involving the S. alba and B. juncea genomes, we recorded approximately 78% genome similarity between the two genera (Singh et al., 2022b). These homologies between the S. alba and B. juncea genomes might have contributed to the successful production of the first stable and fertile somatic hybrids between B. juncea and S. alba, with adequate pairing at meiosis, which maintained the fertility of the hybrids and their backcross progenies (Kumari et al., 2018, 2020c; Kumari and Bhat, 2019, 2021). This has also allowed for the inheritance of the bushy plant architecture and higher numbers of PBs and SBs in the backcross generation (Kumari et al., 2018, 2020b; Singh et al., 2023). This, in turn, has contributed to higher yields.

The identification of S. alba introgressions in advanced backcross generations has been challenging due to the small size of the chromosomes and the presence of more heterochromatic regions, which is similar to the findings in Arabidopsis thaliana (Schweizer et al., 1987). This is because each chromosome in these species has an approximately equal amount of DNA, making it difficult to distinguish segmental introgressions in the recipient genome. Nonetheless, early generations of allohexaploids revealed S. alba introgression via in situ hybridization, which is a powerful technique for visualizing the localization of DNA sequences on chromosomes (Kumari et al., 2020b, 2020c). With the advanced generation of second backcross progeny of somatic hybrids, these BCILs carry enormous allelic diversity due to the presence of half of the haploid set of the S. alba genome, either in the form of segmental or additional chromosomal introgressions. Thus, association analysis by SNP genotyping could not be possible because of sufficient allelic variations within the core set of BCILs. As a consequence, we used a reverse genetics approach in the present study to identify the genes responsible for the greater numbers of PB and SB in a core set of 86 S. alba-B. juncea BCILs by genotyping with a set of 170 S. alba genome-specific SSR markers. However, the chromosome-wide sequence of the S. alba genome is not available to date (Kumari et al., 2020a; Singh et al., 2022b). We used these markers as an alternative approach for establishing a marker–trait association to identify the genes that contribute to the greater number of PB and SB in BCILs, which can contribute to greater yields.

The mixed linear model (MLM) and general linear model (GLM) are statistical methods commonly used in association analysis between genetic markers and phenotypic traits. The MLM approach is preferred due to its ability to control population structure and relatedness among individuals, which can lead to false positives in association analysis. The GLM approach, on the other hand, assumes independence among individuals and can produce spurious associations. In this study, both MLM and GLM were employed, and the Q and K matrices obtained from the population structure analysis and kinship analysis, respectively, were incorporated as covariates to account for the confounding effects of population structure and relatedness (Bradbury et al., 2007). It is interesting to note that despite the small size of the association panel, there was wide variation in branch numbers, indicating the presence of natural genetic diversity. The heritability of the trait was also high, suggesting that genetic factors play a significant role in determining the number of PB and SB. The identification of significant marker–trait associations using the MLM (Q + K) model provides further evidence that genetic factors influence the trait. These associations may help in identifying genomic regions or specific genes that are responsible for regulating branch numbers in BCILs.

Using multiple methods to analyze population structure and determine the most appropriate model for association mapping is a common practice in genetic studies. Population structure is responsible for the identification of numerous false-positive QTLs (Zhao et al., 2007). In this study, the Q+K model and the PCA model were considered, but the MLM (Q + K) was found to be the most appropriate for the population of BCILs being analyzed (Yu et al., 2006; Stich and Melchinger, 2009). Additionally, the use of STRUCTURE analysis, a kinship matrix, and a UPGMA tree helped to categorize the BCILs into three populations, which provided valuable information for association mapping analysis. Overall, it is important to carefully consider population structure and appropriate models when conducting association mapping studies to avoid false-positive results and to accurately identify QTLs associated with traits of interest. Indeed, the modification of plant architecture has been a major focus of crop improvement for several years and has contributed significantly to the increase in crop yield during the Green Revolution (Pingali and Raney, 2005).

Based on the MLM (Q + K) model, a total of five significant marker–trait associations were identified for PB and SB in both crop seasons. The genomic regions associated with these markers contained a total of 47 complete genes, 35 of which were identified as candidate genes for regulating the branching trait in BCILs. These genes were identified as orthologous to AINTEGUMENTA - ANT (Elliott et al., 1996; Mizukami and Fischer, 2000; Krizek et al., 2020), APETALA 1 - AP1 (Irish and Sussex, 1990; Byzova et al., 1999; Monniaux et al., 2018), RAMOSUS– RMS (Arumingtyas et al., 1992; Grbić and Bleecker, 2000; Foo et al., 2005, 2007), REGULATOR OF AXILLARY MERISTEMS - RAX (Keller et al., 2006; Muller et al., 2006), MORE AXILLARY GROWTH - MAX (Stirnberg et al., 2002; Sorefan et al., 2003; Booker et al., 2005; Bennett et al., 2006), MONOPTEROS - MP (Thomas and Gerd, 1993; Przemeck et al., 1996; Christian and Thomas, 1998), SEUSS - SEU (Bao et al., 2010; Gong et al., 2016; Huai et al., 2018), REVOLUTA - REV (Talbert et al., 1995), TERMINAL FLOWER 1 - TFL1 (Bradley et al., 1997), ALTERED MERISTEM PROGRAM 1 - AMP1 (López-García et al., 2016, 2020; Yang et al., 2018), ERECTA - ER (Godiard et al., 2003; Qu et al., 2017), SUPERSHOOT1– SPS1 (Tantikanjana et al., 2001; Ehrenreich et al., 2007), etc. genes that controlling the branching trait. The genes associated with branching traits identified in our study could be utilized in crop improvement programs through targeted genetic modifications, marker-assisted breeding (MAB), and selection (MAS) programs. The introduction of genes responsible for higher branching traits into high-yielding cultivars could lead to increased branching and consequently higher yield potential. This study demonstrated the introgression of 47 genes into B. juncea from S. alba and identified 35 of these genes as candidates involved in increased PB and SB traits (Ehrenreich et al., 2007). The identified genes were subjected to BLASTN analysis, which matched their sequences with those of the transcriptome assembly of the first stable allohexaploid Brassica (H1) strain developed from RNA-seq at the time of flowering (Singh et al., 2022a). This study highlights the potential for gene introgression to improve crop traits, such as branching, through the transfer of genetic material from related wild species.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

KS: Data curation, Investigation, Methodology, Software, Supervision, Validation, Writing – original draft. PK: Conceptualization, Data curation, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing. PR: Resources, Writing – review & editing.