Roots are a fundamental component of the plant vascular system, playing a pivotal role in plant growth, development, and overall survival. The primary function of roots is absorbing water and nutrients from the soil, as well as serving as an anchor that firmly secures plants to the ground, thereby ensuring stability and support. The development of a root system is an important quantitative characteristic that determines a plant’s capacity to survive across different environments. Improved understanding of the behavior of roots within natural ecosystems is of significance in enhancing crop yields, developing more resilient plant varieties, and preserving biodiversity (Griffiths et al., 2022).

The genus Brassica consists of 37 species including the widely cultivated commercial crops Brassica napus L. (AACC, n = 19), B. rapa L. (AA, n = 10), B. nigra (L.) Koch (BB, n = 8), B. oleracea L. (CC, n = 9), B. juncea (L.) Czern & Coss (AABB, n = 18), and B. carinata A. Braun (BBCC, n = 17) (Branca and Cartea, 2011). Most B. juncea and B. nigra genotypes have fibrous roots (Center for Agriculture, Food, and the Environment, 2015), while B. napus (Arif et al., 2019), B. rapa (Center for Agriculture, Food, and the Environment, 2015), B. oleracea (Center for Agriculture, Food, and the Environment, 2015), and B. carinata (Barro and Martín, 1999) have a large taproot system with a single main root axis and hundreds of lateral roots. Several root system architecture (RSA) traits are heritable characteristics, which suggests some level of genetic control mechanisms that may be exploited in breeding programs (Shi et al., 2013). In addition, root growth and development have been reported to be influenced by plant hormonal cues (Tanimoto, 2005; Santner et al., 2009; Vissenberg et al., 2020). The uneven distribution of nutrients and water in soils as well as soil microbiomes have also been reported to influence root growth and development (Molefe et al., 2023). Lastly, the plasticity of roots make them vulnerable to physico-chemical environmental influences (Lynch, 2007; Ingram and Malamy, 2010; Yang et al., 2010; Shi et al., 2013; Zhang et al., 2016; Ibrahim et al., 2021).

Genetic variability in plants is a fundamental component of biodiversity and serves as the foundation for the creation of novel and enhanced cultivars with desirable traits (Govindaraj et al., 2015; Yu et al., 2021). Therefore, increasing the genetic diversity in Brassica genotypes is an important breeding objective. However, only a few studies have reported the potential of root phenomics for crop breeding (Kuijken et al., 2015; Prince et al., 2019; Falk et al., 2020; Liu et al., 2021). This is mostly due to the complexity and large genetic variability of Brassica root systems, the difficulties of accessing intact roots, as well as the tedious and laborious nature of phenotyping Brassica roots (de Dorlodot et al., 2007; Zhu et al., 2011; Meister et al., 2014; Ibrahim et al., 2021; Yang et al., 2024). The main goal of breeding programs that target root traits is to enhance the ability of roots to explore the soil and acquire water and nutrients, and to develop crops with increased stress tolerance and improved yields. By improving understanding of RSA and the underlying pathways that shape it, researchers can leverage diverse root features to help plants respond to climate change and enhance crop yields (Smith and De Smet, 2012; Ibrahim et al., 2021). However, genetic studies focusing on RSA, especially of Brassica species, are limited compared with research on aboveground traits. Most of these studies were conducted in controlled environments, but have value since they enable the observation of RSA and the collection of intact roots (Wang et al., 2019). For example, the detection of overlapping quantitative trait loci (QTLs) for root traits under non-soil-based conditions alongside those influencing yield in the field suggested the relevance and applicability of the non-soil methods (Tuberosa et al., 2002; Wang et al., 2019).

The limited genetic studies on Brassica RSA have focused primarily on the important role of root morphology in enhancing the efficiency of phosphorus, boron, and nitrogen uptake in rapeseed (B. napus) (Yang et al., 2010; Shi et al., 2013; Zhang et al., 2016; Wang et al., 2017, 2019). Alcock et al. (2018) reported that in B. napus, the chromosomal regions associated with foliar phosphorus concentration harbor multiple genes that influence root development. Furthermore, homologs of three ABC transporter proteins implicated in root suberin synthesis exhibited co-localization with peaks associated with leaf nitrate and phosphorus levels. In another RSA study, Ahmad et al. (2023) identified 39 significant single nucleotide polymorphisms (SNPs) on Brassica chromosomes A04, A09, A10, C02, C03, C07, and C09 that were associated with RSA and biomass traits in 327 rapeseed accessions under low phosphorus conditions. Fletcher et al. (2016) also reported 20 QTLs on chromosomes A03, A10 and C02, associated with flowering time and root mass in a B. napus doubled haploid population, demonstrating their association with adaptation to drought.

Recently, we identified total root length (TRL/cm), total root surface area (TRSA/cm²), root average diameter (RAD/cm), number of tips (NTP), total primary root length (TPRL/cm), total lateral root length (TLRL/cm), total tertiary root length (TTRL/cm), and basal link length (BLL/cm) as significant RSA traits in the B. napus root system (Yang et al., 2024). The genomic regions controlling the observed phenotypic differences, however, were not determined. Genome-wide association studies (GWAS) is a method rooted in the concept of linkage disequilibrium (LD), which detects associations between genotypic and measured phenotypic data. This is achieved by examining differences in allele frequency of genetic variants within natural populations (Uffelmann et al., 2021). This approach offers notable advantages compared with traditional linkage-based association mapping (Gupta, 2016; Fredua-Agyeman et al., 2020). GWAS enable the exploration of a broader range of allelic diversity, providing enhanced resolution for analyzing various traits of interest. In addition, GWAS also offer the opportunity to examine genotypes across different crop species, eliminating the need for ancestry or pedigree data typically required in QTL mapping (Fredua-Agyeman et al., 2020).

In the present study, we assessed eight RSA traits in 379 accessions representing the six commercially important Brassica species, B. napus, B. juncea, B. carinata, B. oleracea, B. nigra, and B. rapa. Additionally, we identified the genomic regions associated with these key RSA traits through GWAS.

Root system architectural traits are critical to the plant’s ability to absorb water and nutrients from the soil (Zobel et al., 2007; Lynch, 2019; Wen et al., 2019; Liu et al., 2021; Sun et al., 2021). The response of crops to abiotic stresses is influenced by their RSA. Since roots grow underground, they serve as the first line of defense in detecting stress signals and adapting their genetic program for post-embryonic growth to cope effectively with these challenges (Lynch, 1995). Morphological variations often correspond to physiological or functional variations (de Dorlodot et al., 2007). In this study, a highly positive correlation was observed for seven of the eight RSA traits (except RAD) evaluated in the 379 Brassica genotypes. Ibrahim et al. (2021) also reported a positive correlation between RSA traits in a study of 388 B. napus accessions. The mean BLL values were similar across all six species examined, indicating that the potential growth of the roots of these species is quite similar, probably due to the phenotyping platform.

Root architectural traits can serve as a focal point for enhancement and optimization, tailored to meet the specific requirements dictated by soil conditions (Arifuzzaman and Rahman, 2017). For example, Thomas et al. (2016) examined the root structure of various growth types of B. napus and determined that root morphology has the potential to enhance crop yield, provided that appropriate genetic markers associated with agronomic traits can be identified (Arifuzzaman and Rahman, 2017). The findings from this study revealed that root traits dynamics were significantly influenced by the genotype and species, highlighting the significant role of genetic factors in RSA. The presence of shared QTL between root characteristics and productivity measures such as yield, water usage, or nutrient acquisition suggests that the former contributes to determining the latter in numerous instances (Steele et al., 2006; Ibrahim et al., 2021). One example is the utilization of a QTL termed DRO1, which governs both root growth angle and root depth in rice, for improving the root traits of an Indian upland rice variety through marker-assisted selection to enhance water efficiency (Steele et al., 2006; Ibrahim et al., 2021). In the case of Brassica crops, additional studies regarding the relationship between root-related traits and productivity or nutrient and water use efficiency are needed.

From a functional perspective, a root tip is the primary site for nutrient and water uptake in a plant. In addition, root tips produce exudates that attract significantly higher levels of microorganisms to the rhizosphere surrounding the root tips compared with the bulk soil located a few millimeters away (Curlango-Rivera and Hawes, 2011). Furthermore, individual root tips that develop in soil provide strong anchorage, which facilitates deeper penetration of roots into hard soil niches or areas (Bengough et al., 2016). In this study, significantly greater NTP per plant in B. napus and B. oleracea accessions suggests they possess the potential for stronger soil anchorage, as well as better nutrient and water uptake capabilities. Longer primary roots (Wasson et al., 2012), a larger root diameter (Uga et al., 2013), and abundant and steeper lateral roots (Lynch, 2013) were reported to lead to a deeper and more resilient root system with increased radial hydraulic conductivity at depth and decreased metabolic costs for drought adaptation (Khan et al., 2016). The aforementioned root traits were also reported to increase the efficiency of exudation of organic anions (Lynch, 2015) and enhance interactions with microbes (Walch-Liu et al., 2006), resulting in a high tolerance to deficiencies in nutrients such as nitrogen and phosphate (Khan et al., 2016). In the current study, the B. napus and B. oleracea accessions that possessed relatively larger root systems with larger root surface area, and longer and more vigorous roots, would be expected to provide good anchorage and penetration into the soil. This suggests high developmental plasticity in case of drought or nutrient deficiency (Choi and Cho, 2019). Akhatar and Banga (2015) reported a positive association between seed yield and root length in B. juncea under irrigated conditions. However, negative correlations between root length, root surface area, and root mass with yield under drought stress have been observed in potatoes (Solanum tuberosum) (Zarzyńska et al., 2017). Hammond et al. (2009) and Pongrac et al. (2020) reported that genotypes of B. oleracea with high phosphorus absorption efficiency were characterized by more and longer lateral roots, and had significantly increased yields independent of external phosphorus concentration. Thus, the B. oleracea genotypes with large TLRL identified in this study will have the potential for high phosphorus absorption efficiency, and likely have the potential for higher yields. However, the efficiency of nutrient use is complex and influenced by physiological traits specific to each species and other factors (Pongrac et al., 2020).

The distribution of roots, especially those with strong and thick root systems capable of deeper penetration in the soil, is pivotal in influencing a plant’s capacity to acquire essential resources like water (Fenta et al., 2014). Compared with other species, B. juncea and B. carinata, with the largest RAD, might be better at surviving in dry and compacted soil given their relatively thick and stronger roots. Similar patterns have been identified in other crop species. For example, rice (Oryza sativa L.) varieties with deeper and thicker root systems exhibited increased yields and quality under drought conditions (Suji et al., 2012). Similar trends were also observed in four common bean (Phaseolus vulgaris L.) cultivars, wherein varieties with deeper and stronger roots exhibited enhanced yield, growth, reduced canopy temperature, and decreased soil moisture extraction; in poor soil, however, root variations were not evident (Sponchiado et al., 1989; Lopes et al., 2011).

Salinity is an abiotic stress with detrimental effects on agricultural productivity and sustainability, making it a significant global concern (Shahid et al., 2018). Approximately 830 million ha are projected to be affected by salinity (Szabolcs, 1989), which is still increasing every year (Qadir et al., 2014; Minhas et al., 2020). Agricultural productivity under saline conditions is low, posing a perpetual risk of crop failure. This underscores the significance of enhancing crops to thrive in such environments (Minhas et al., 2020). Reducing main root elongation limits the transport of sodium ions from roots to shoots in saline soils (Rus et al., 2006; Munns and Tester, 2008; Katori et al., 2010). In addition, the storage of sodium ions in root vacuoles and steles can increase the efficiency of water extraction and ion exclusion for salinity tolerance (Gupta and Huang, 2014; Khan et al., 2016). The current study suggests that B. nigra and B. rapa, which had relatively shorter primary roots compared with the other four species, may exhibit greater tolerance to salinity.

Soilborne pathogens also represent a major threat to crop production. Clubroot, caused by Plasmodiophora brassicae, is a particularly important disease of Brassica species (Strelkov and Hwang, 2014). In general, Fredua-Agyeman et al. (2019) found that the order of clubroot resistance, from greatest to smallest, was B. nigra > B. oleracea > B. rapa > B. napus. As such, while B. nigra had the smallest root system in this study, this species also appears to have the strongest resistance to clubroot. A similar association was noted between root size and Verticillium stripe (Verticillium longisporum) severity (Cui et al., 2023). Inoculation of 3-week-old canola plants resulted in more severe disease compared with inoculation of 1- or 2-week-old plants, which had smaller root systems. Furthermore, B. nigra, B. carinata, and B. juncea, characterized by relatively smaller and less intricate root systems in the present study, also exhibited greater resistance to Sclerotinia stem rot (Sclerotinia sclerotiorum) (Navabi et al., 2010; Neik et al., 2017) and blackleg (Leptosphaeria maculans and L. biglobosa) (Roy, 1984; Sjödin and Glimelius, 1988; Rimmer and van den Berg, 1992; Li et al., 2015; You et al., 2016; Neik et al., 2017). Infection by soilborne pathogens can destroy the roots, reduce root density, and diminish the functional effectiveness of the surviving infected roots (Román-Avilés et al., 2004). It is likely that host genotypes with smaller root systems offer fewer opportunities for pathogen invasion, given the smaller surface area and reduced opportunity for contact. As such, the greater resistance may be due to disease escape during the early growth stages of the plant. In the GWAS, 6,213 SNP markers were used to measure RSA traits, including 5,103 A-genome and 1,110 C-genome markers. Comparative genomic studies on Brassica genomes have reported that 1 centimorgan (cM) on a genetic map corresponds to ~500 kb (Suwabe et al., 2006; Ecke et al., 2010; Delourme et al., 2013). Therefore, the 302.5 Mb marker coverage estimated in this study for the A-genome and 452.8 Mb for the C-genome corresponded to ~605 cM and ~905 cM, respectively. As such, the 6,213 SNP markers covered a total of ~1,510 cM, which represents ~60% of the estimated 2,500 cM B. napus genome. The determined genome coverage was comparable with a value of ~645 Mb obtained in studies using a Brassica 60K array (Qian et al., 2014; Qu et al., 2017) and specific-locus amplified fragment sequencing technology (Zhou et al., 2017). In comparison, the filtered set of 6,213 markers on the Brassica 19K SNP array provided about 3× more coverage than the Brassica 13.2K SNP array from the same company used in a previous study (Fredua-Agyeman et al., 2020).

The mean marker density using the Brassica 19K SNP array was 62.9 ± 20.1 (8.43 SNP markers/cM) on the A-genome, 426.3 ± 106.9 (1.22 SNP markers/cM) on the C-genome, and 235.0 ± 200.1 (4.1 SNP markers/cM) on the entire genome. In comparison, the mean marker density using the Brassica 13.2K SNP array was 63.4 ± 21.9 (8.46 SNP markers/cM) for the A-genome, 15.0 ± 8.4 (44.3 SNP markers/cM) for the C-genome, and 40.5 ± 29.8 (11.8 SNP markers/cM) for the entire B. napus genome (Fredua-Agyeman et al., 2020). Thus, the marker density remained consistent for the A-genome, but it was 2–3× lower on the C-genome when genotyping was conducted with the Brassica 19K vs. 13.2K SNP array. This discrepancy was anticipated, as the 1,110 filtered set of SNP markers on the C-genome was distributed over 302.5 Mb or 905.5 cM on the 19K array, in contrast to 2,367 markers on the C-genome spread over 26.7 Mb or 53.4 cM on the 13.2K array.

Linkage disequilibrium, which is the non-random association between alleles at different loci, determines the power and precision of association mapping studies using molecular markers and unobserved QTL (Goddard and Hayes, 2009; Qu et al., 2020). Determination of the extent of LD is essential for making inferences regarding the genetic forces shaping a population (Qanbari, 2019). The extent of LD reported by Fredua-Agyeman et al. (2020) using the Brassica 13.2K SNP array varied from 1,100 to 2,300 kb for the A-genome and from 200 to 1,500 kb for the C-genome. In the present study using the Brassica 19K SNP array, LD varied from 440 to 1,400 kb for the A-genome and from 2,500 to 9,100 kb for the C-genome. The difference in LD values obtained with the two Brassica arrays could reflect the different marker densities. The low marker density on the C-genome might be responsible for the extended ranges of the LD decay. However, the LD values for the A- and C-genomes were consistent with those reported in other studies (Wu et al., 2016; Qu et al., 2017; Zhou et al., 2017). Based on the minimum LD decay (440 kb or 0.88 cM), a minimum of 3,200 markers were needed to perform the GWAS. Therefore, the 6,213 SNP markers in this study represented approximately twice the number needed to perform the analysis.

The MTAs analyses identified three genomic regions on chromosomes A02, A03, and A06 that were associated with RSA traits. In the case of the A02 chromosome, the SNP marker Bn_A02_p5571981 overlapped with a histone deacetylase HDT2-like protein. This protein negatively regulates GIBBERELLIN 2-OXIDASE2 (GA2ox2) expression, which determines cell number in the Arabidopsis root meristem and elongation zone (Li et al., 2017). The increased expression of GA2ox2 in HDT1/2 was reported to cause a decrease in gibberellin (GAs) levels, leading to an earlier transition from cell division to the expansion phase of transit-amplifying cells (Li et al., 2017). On chromosome A03, a histidine kinase 2 (AHK2) encoded by genes that overlap SNP marker Bn_A03_p19974784 can positively regulate the level of cytokinin, which negatively regulates root development in Arabidopsis (Nishimura et al., 2004; Riefler et al., 2006). In Arabidopsis, histidine kinase homologs function as receptors for cytokinin and play an overlapping role in regulating the growth of shoots and roots (Nishimura et al., 2004). Root hair specific 17, encoded by genes associated with SNP marker Bn_A06_p26219274 on chromosome A06, is an expressed protein controlling root hair cell expression for regulating the root growth of Arabidopsis (Won et al., 2009).

This study characterized the RSA traits of the six most economically important Brassica species. It represents the most comprehensive study of its kind, where the root traits of such a large number of Brassica accessions have been examined. This broader scope enabled a deeper exploration of the genetic basis underlying root development, and its implications for crop improvement strategies. However, this study was conducted in a hydroponic system, which does not fully reflect root growth in soil environments. Therefore, further studies will focus on cultivating a subset of identified accessions with small, medium, and large root sizes in soil to validate the findings. Additionally, accurate mapping of genomic regions associated with RSA traits in Brassica species is necessary. To achieve this, bi-parental mapping populations (F₂ and doubled haploid) will be developed through crosses between selected large and small root-sized accessions to precisely map these genomic regions. Despite these limitations, this study, which encompassed a wide range of Brassica species and root traits, has laid the groundwork for future investigations aimed at breeding programs tailored to enhance root traits and stress tolerance in Brassica crops.

Overall, this study demonstrated that significant genetic variation exists in the RSA of different Brassica species under controlled environmental conditions. The results also indicated correlations between specific RSA traits, with TRL, TRSA, TPRL, and TLRL exhibiting the strongest associations. These identified RSA traits can serve as valuable indicators for further investigations into stress tolerance under field conditions or abiotic stress scenarios. GWAS identified significant MTAs associated with proteins involved in cell wall synthesis, cell growth, and organ morphogenesis, as well as various proteins associated with cellular, biological, and physiological processes involved in root growth and development. The candidate genes related to root growth were located on chromosomes A02, A03, and A06 of the B. rapa genome. We identified B. napus, B. juncea, B. rapa, B. nigra, B. oleracea, and B. carinata accessions with variable RSA traits. These accessions hold promise for breeding Brassica crops suitable for different environments.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .

CY: Conceptualization, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft. RF: Conceptualization, Formal analysis, Methodology, Resources, Software, Supervision, Writing – review & editing. SH: Conceptualization, Data curation, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing. LG: Methodology, Resources, Software, Writing – review & editing. SS: Conceptualization, Data curation, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing.