The sorghum-sudangrass hybrid (Sorghum bicolor × Sorghum sudanense [Piper] Stapf) is an annual warm-season forage that provides water and soil conservation benefits. It combines the advantages of high grass and large leaf yield, improved tillering, good regeneration, and high nutritional value from both parents, providing substantial heterosis. Cattle, sheep, geese, ducks, poultry, and other livestock utilize this hybrid as feed. The sorghum-sudangrass hybrid presents a developed root system, good soil capacity, and strong ecological adaptability and further presents broad development and utilization prospects in environmental protection and animal husbandry, especially the development of the breeding industry in agricultural areas (Saballos et al., 2008; Han et al., 2015; Mahmoudzadeh and Oad, 2018; Peng et al., 2020). In recent years, with the rapid development of animal husbandry and frequent extreme weather worldwide, increasing demand for high-quality forage has indicated a shortage of high-quality green and fresh forage. This is the main reason for limiting the quality and efficiency of animal husbandry in the northwestern Himalayan region of India (Mir et al., 2020; Kumar et al., 2022), the southern part of the Korean Peninsula and Jeju Island (Peng et al., 2020), the northern and southeastern United States (Wells et al., 2014; Nave et al., 2020), northern Europe (Liatukienė et al., 2008), Canada (Bélanger et al., 2006), and parts of China (Fang, 2018). Therefore, increasing forage yield is crucial in meeting sufficient livestock nutrition and accelerating production. Recently, the sorghum-sudangrass hybrid has been widely cultivated and applied due to its high grass yield, strong resistance, rich nutrient content, and wide adaptability, which will ultimately play a positive role in sustainable animal husbandry development and alleviate the contradiction between supply and demand of herbage.

Forage grass is a crucial material base of animal husbandry and the main source of livestock nutrition (Li et al., 2015; Zhang et al., 2021). Plant height (PH), stem thickness (ST), tiller number (TN), leaf-related traits (leaf number (LN), leaf length (LL), and leaf width (LW), and fresh weight (FW) represent the most important yield traits. An understanding of the genetic basis of these traits carries considerable importance regarding increasing the output of sorghum-sudangrass hybrids (Ma et al., 2018). In the past decades, gene identification related to crop target traits was mainly achieved via hybrid population breeding (Girma et al., 2019). Researchers from the United States, India, Japan, Australia, and China have cultivated several forage varieties, such as Jianlibao (Jumbo), Leyi (Everlush), Mengong Green series forage, Wancao No. 2 (WC-2), and Jicao No. 1 (JC-1) (Han et al., 2015). However, conventional breeding results from the perennial selection of crop hybrid populations for important phenotypic traits, which are time-consuming and inefficient regarding targeted modifications for complex traits (Huang and Yan, 2019; Islam et al., 2021). With molecular biology’s rapid development, molecular marker-assisted breeding has received increasing attention in forage breeding. Understanding the quantitative trait locus (QTL) function associated with important yield traits is an effective way to improve crop yield.

QTL mapping has been widely used to study yield traits in sorghum-sudangrass hybrids. Shi et al. (2017) constructed genetic linkage maps based on amplified fragment length polymorphism (AFLP) molecular markers and detected 13, 11, 10, 2, and 9 QTLs for stem diameter (SD), LL, TN, LW, and LN, respectively. Wang et al. (2021) detected four QTLs associated with PH and one QTL related to TN based on a genetic map containing 133 SSR markers. Using a genetic linkage map containing 181 SSR markers, Yu et al. (2018) further detected two QTLs related to LN, LL, LW, and TN each.; four QTLs with SD; and one QTL associated with PH. Liu et al. (2015) constructed a genetic linkage map with 124 SSR markers and identified one QTL each for PH, SD and TN, and three QTLs each for LN and FW. However, genetic maps constructed using traditional molecular markers, such as simple-sequence repeats (SSR) and AFLP, have low density and few QTL sites, which complicates meeting the needs of further QTL fine mapping (West et al., 2006).

With the gradual increase in available plant genome sequences, the combination of next-generation sequencing technology and reference genomes provides a new research strategy for QTL mapping and molecular marker-assisted breeding of important plant traits (Hart et al., 2001). The genome size of sorghum is about 730 Mb. Since the sudangrass and sorghum-sudangrass hybrid genome sequencing was incomplete, we used the sorghum genome as a reference genome to develop single nucleotide polymorphism (SNP) markers for sorghum-sudangrass hybrid. Compared to AFLP and SSR markers, SNP markers are abundant and polymorphic, allowing high-density genetic map construction (Wen et al., 2020). A single genetic linkage map has been constructed using SNP markers in sorghum-sudangrass hybrids (Jin et al., 2021). SNP markers can be developed on a large scale through different high-throughput sequencing technologies, such as genotyping-by-sequencing (GBS) (Elshire et al., 2011), restriction site-related DNA tag sequencing (RAD-seq) (Xu et al., 2018), specific-length amplified fragment sequencing (SLAF-seq) (Sun et al., 2013), and whole-genome resequencing (WGRS) (Li et al., 2009). WGRS is a sequence-based genotyping method that is not limited by a restriction site and can quickly obtain a considerable number of recombinant breakpoints and marker density. According to research needs, sequencing coverage can be adjusted to the whole or part of the population, which vastly improves the accuracy of QTL mapping (Hart et al., 2001). WGRS has been successfully applied to map construction and QTL mapping of related traits of a variety of important crops, such as rice (Smulders et al., 2019), sorghum (Zhang et al., 2021), and peanuts (Jiang et al., 2021). However, WGRS has not been used in QTL mapping for traits related to the sorghum-sudangrass hybrid, and it is rarely reported in gene prediction.

Therefore, the aim of this study was to conduct QTL mapping analysis for seven traits (i.e., PH, SD, TN, FW, and the leaf-related traits LN, LL, and LW), and the genes in the stably detected QTL intervals were annotated based on the high-density SNP map of sorghum-sudangrass hybrids constructed in previous research by our group. These results will serve as a foundation for further studies such as QTL fine mapping, functional key gene verification, and marker-assisted breeding.

The rapid development of molecular biology and the wide application of high-throughput sequencing technology have promoted new breeding strategies to increase crop yield and improve important yield-related traits. Several genetic linkage maps have been created, and some progress has been made in the QTL mapping of related traits. However, most of the existing genetic maps were constructed using restriction fragment length polymorphism (RFLP), sequence-related amplified polymorphism (SRAP), and simple-sequence repeats (SSR), which present few markers and large QTL confidence intervals, limiting their use in QTL fine mapping and marker-assisted breeding (Liu et al., 2015; Shi et al., 2017; Yu et al., 2018; Wang et al., 2021). Compared with the above molecular markers, SNP markers are widely used in map construction and QTL mapping of a variety of crops due to their high density, uniform and extensive distribution on chromosomes, and high genetic stability (Smulders et al., 2019; Jiang et al., 2021; Zhang et al., 2021). Our research group previously constructed a high-density genetic linkage map of the sorghum-sudangrass hybrid, which contained 5656 SNP markers, covering a total genome length of 2192.84 cM, and the average distance between markers was 0.39 cM. (Lu et al., 2022) Compared with the previously constructed maps, the marker density increased, which effectively improved the accuracy of QTL mapping and provided a possibility for further screening of key genes and fine mapping of QTLs.

With the rapid development of sequencing technology and cost reductions, an increasing number of crop genomes have been exploited and applied in related studies (Li et al., 2021). Methods to detect QTL loci physically similar to or overlapping with the reference genome by combining specific crop genome sequences (at the chromosomal level) and high-density genetic maps have been widely used in the rapid identification of QTL loci and potential candidate genes (Luo et al., 2020). Since the sequencing of sudangrass has not been completed, Jin et al. (2021) reported that it is feasible to use the sorghum genome as the reference genome of the sorghum-sudangrass hybrid. In this study, 55 QTLs related to yield traits were detected in three environments based on a high-density genetic linkage map ( Table 2 ). Markers located in the same or similar positions on the same chromosome as those in previous studies are listed in Table 4 . qFW-6-1 is a stable QTL associated with FW and was detected in two of the three environments. In contrast to the previous studies, qFW-6-1 (2402139–50582068 bp) was close to QFBMS6.1 (47686626–50991177 bp) (Wang et al., 2014), qFW6 (45, 156, 899–55, 463, 230 bp) (Jin et al., 2021), and qTW6 (49894350–51216671 bp) (Kajiya-K et al., 2020), which explained 21.22–21.39% of the phenotypic variation, and is an important QTL associated with FW. The marker qLW-4-4 (170.61 cM) was close to the P9m58-453-P9m58-208 marker (171.1 cM) (Shi et al., 2017) (with a negligible difference of 0.49 cM between the two sites), and the phenotypic variation was as high as 25.95%, which was the main effect site of LW. Furthermore, qLN-5-1 is located at 20.41 cM (19.8–22.3 cM) on LG 5, close to marker LN2-1 (22.9 cM) (Yu et al., 2018), with a phenotypic variation of 14.37%. It is a major QTL associated with LN. In addition, most QTLs located on the same LG were detected in the new marker interval due to the differences in marker types, population types, size, and material planting environments used to construct genetic maps (Guo et al., 2020). For PH, two stable QTLs were identified, namely qPH-7-1 and qPH-7-2, which were located at 4.71 cM and 8.11 cM on LG 7. The qPH-7-1 and qPH-7-2 LOD values were 3.4–5.2 and 3.0–4.4, respectively, and the additive effect between alleles was positive, which explained 7.76–11.66% of the phenotypic variation, indicating a stable QTL locus associated with PH. Compared with previous studies, qPH-7-1 (4.2–5.4 cM) and qPH-7-2 (7–10.5 cM) were co-localized with qPH7 (110.3–112.92 cM) loci on chromosome 7 (Zou et al., 2012). However, both qPH-7-1 and qPH-7-2 were inconsistent with qPH7 markers, suggesting that these may be novel QTLs for PH. In subsequent studies, these QTLs will be the focus of our attention. Additionally, compared with previous studies, it was found that several QTLs related to yield traits were detected at different positions on the same chromosome. For example, the QTLs controlling SD was detected on LG7, LG8, LG9, and LG10 (Wang et al., 2014; Liu et al., 2015; Kong et al., 2020; Jin et al., 2021); that controlling LL on LG3, LG4, LG7, LG8, LG9, and LG10 (Wang et al., 2014; Shi et al., 2017; Yu et al., 2018); and the one related to TN on LG2 (Li et al., 2015; Rayaprolu et al., 2021).

Yield traits such as PH, SD, TN, leaf-related traits, and FW are important quantitative traits, which are influenced considerably by the environment. The results of QTL localization are different in different environments, and the accuracy of the localization results can be guaranteed by setting up multi-year and multi-environmental tests to verify the QTL detected in multiple environments (Wang et al., 2009; Hou et al., 2015). Generally, QTLs that can be located in multiple environments (at least two environments) or all environments and have similar effects are defined as stable QTLs; otherwise, they are considered to be greatly influenced by the environments that they interact with (Xie et al., 2008). Guo et al. (2020) detected 38 QTLs related to protein and 68 related to starch in three environments and the average environment, respectively, and 26 stable QTLs were detected in more than two environments. Liu et al. (2022) identified 183 QTLs related to cotton fiber and yield traits in six environments, 62 QTLs for fiber and 10 QTLs for yield stability were identified in multiple environments. Ma et al. (2018) mapped maize leaf-related traits in three environments and eight stable QTLs in two or three environments, explaining 4.38–19.99% of the phenotypic variation. Additionally, Yang et al. (2022) identified 105 QTLs related to cotton in three environments, a total of 25 stable QTLs were detected in more than two environments. In this study, 55 QTLs related to yield traits of the sorghum-sudangrass hybrid were mapped in 3 environments and the average environment, and 17 RHF-QTLs were repeatedly detected in at least two environments ( Table 3 ). Among them, there were two stable sites related to PH, four in SD, one in LL, two in LW, three in LN, and five in FW. These could explain 4.9–30.97% of the phenotypic variation. qPH-7-1 and qPH-7-2, which control plant height, could be detected in 2021-H, 2022-T, and the average environments, explaining 7.76–13.56% of the phenotypic variation (>10%), which may be stable QTL controlling PH formation. Meanwhile, we also found that a pair of stable QTLs controlling different traits were co-localized at the same position and chromosome. The qLL-10-1 (23–26.6 cM, 25.81 cM) and qFW-10-4 (23–29.6 cM, 25.81 cM) markers were completely or only partially overlapped in LG10.

The LRR receptor-like serine/threonine-protein kinase encoded by the gene23531 has high homology with the brassinolide insensitivity (BRI) gene and jointly regulates brassinosteroid signals. Ou et al. (2015) investigated dwarfing and non-dwarfing rootstock pear varieties based on RNA-sequence, and the results showed that the LRR receptor-like serine/threonine-protein kinase, a key gene controlling PH growth, was significantly up-regulated in the dwarfing varieties (Li et al., 2002). The psbP domain-containing protein encoded by gene26589 participates in plant photosynthesis and plays an important role in the assembly of plant PS II and maintaining conformation stability (Bricker et al., 2013). Many studies have shown that the suppression of PsbP will cause the decline of the oxygen evolution ability of plants, a change in the direction of the electron transfer chain, and a lack of PS II oxidation and reduction function (Yi et al., 2007; Ido et al., 2009). The gene15585, coding the galacturonosyltransferase 8 (GT8) family proteins, is divided into two branches. The former mainly includes subclades of galacturonosyltransferase (GAUT) and galacturonosyltransferase-like TL (GA) genes, which play an important role in the synthesis of cell walls (Cheng et al., 2018). Kong et al. (2011) found that GAUT1 was involved in the synthesis process of pectin. In a tomato study, de Godoy et al. (2013) found that GAUT4 gene silencing treatment significantly reduced the pectin content. Therefore, it was speculated that gene15585 might be key in leaf growth. The magnesium transporter NIPA2, encoded by the gene15584, is involved in plant photosynthesis and plays an important role in leaf growth and aging delay. Hermans and Verbruggen (2005) and Horlitz and Klaff (2000) also found that magnesium transporter NIPA2 is an important Mg²⁺ transporter, which could accelerate the continuous transport of Mg²⁺ to green tissues such as leaves, promoting the synthesis of green pigment and carbon assimilation and accelerating plant growth. It was speculated that the gene might be related to leaf photosynthesis. Chloroplasts are semi-autonomous organelles that contain their own DNA and can self-replicate. The gene23381 and gene24523 encode chloroplast DNA polymerase and RNA polymerase, respectively, which are key enzymes in plant regulation of chloroplast DNA synthesis and transcription in plants. RNA polymerase participates in the synthesis of various mRNA species, transcribes tRNA genes, regulates rRNA synthesis, and plays an important role in maintaining the growth and development of plant leaves (Börner et al., 2015). Zoschke et al. (2007) measured chloroplasts in Arabidopsis seeds and young, and old leaves at the transcriptional level and found that the transcriptional activity was relatively stable in all three. In this study, gene23381 and gene24523 were annotated in qLN-5-1, a QTL related to LN traits, suggesting that these genes may be important in controlling the growth and development of leaves of the sorghum-sudangrass hybrid. The gene31524 encodes plant cytochrome P450. Renault et al. (2014) found that it has high catalytic activity, participates in a variety of metabolic reactions in plants, and plays an important role in signal transduction, pigment synthesis, light, electron transport, and biological defense. Therefore, it is speculated that gene31524 may be the key gene affecting the growth and metabolism of the sorghum-sudangrass hybrid.

Under different environmental conditions, QTL loci detected at the same or adjacent loci on the same chromosome are called “QTL hotspots” or “QTL clusters,” which are the result of the regionalized distribution of QTLs related to different traits (Yang et al., 2022). They are also the preferred regions for fine mapping and candidate gene identification (Xie et al., 2008). It can introduce genes related to crop quality, yield, and resistance into recipient crops simultaneously, control the correlation between traits in stable “QTL hotspots” or “QTL clusters” of different traits, and regulate pleiotropy of various traits through different metabolic pathways (Huang and Yan, 2019; Waheed et al., 2021). Numerous studies have shown that the phenomenon of “QTL hotspots” or “QTL clusters” are prevalent in a variety of crops, such as crested wheatgrass (Yang et al., 2022), wheat (Cui et al., 2016), and sorghum (Mace et al., 2012). In this study, we found that qLL-10-1 and qFW-10-4 were located at the same position of 25.81 cM on LG10 among the seven stable QTL controlling yield traits of sorghum-sudangrass hybrid. In addition, FW was significantly positively correlated with LL, in agreement with previous studies.

In this study, 55 major QTLs related to PH, SD, TN, LL, LW, LN, and FW were identified based on the high-density SNP map of the sorghum-sudangrass hybrid, among which 17 relatively RHF-QTL were detected in at least two environments. A stable QTL cluster containing QTLs controlling LL and FW (including at least one RHF-QTL) was detected, and three QTLs overlapping or located adjacent to the previously studied sites were identified. The genes in the RHF-QTL intervals were annotated, and seven candidate genes that might be related to PH, SD, LW, LN, and FW were screened. The results of this study will promote the fine mapping of QTL for yield traits of sorghum-sudangrass hybrids, cloning of key genes, and marker-assisted breeding.

The data presented in the study are deposited in the figshare repository, accession number https://doi.org/10.6084/m9.figshare.21717146.v1.

ZY and QL conceived and designed the study. ZY, XY and QL performed the experiments. QL and XY wrote the article. HW, XZ, and YZ assisted in the performance of the experiments. ZY is the corresponding authors at the request of the Institute. All authors have read and approved the final manuscript.