Sorghum can grow on marginal lands due to its remarkable tolerance to various abiotic stresses such as drought, high salinity, and low nutrient levels (Figure 2) (Berenji et al., 2011; Maharajan et al., 2021; Woldesemayat and Ntwasa, 2018). Sorghum primarily grows in arid regions of Asia and Africa, known as the “camel of crops” for its ability to thrive in drought-stricken soils and endure prolonged droughts (Assefa et al., 2010; Hariprasanna and Patil, 2015). Several potential sources of stay-green characteristics have been identified in sorghum, contributing to yield under both well-watered and drought conditions (Borrell and Hammer, 2000; Jordan et al., 2012). As a C₄ plant, Sorghum has over 30% greater water-use efficiency than C₃ crops (FAO, 2020). This allows it to thrive in areas with less than 400 mm of annual rainfall, supporting farmers in resource-limited regions. Potential traits, associated with tolerance to heat stress, have been found in sorghum, such as early morning flowering and reduced canopy temperature. The early morning flowering trait allows sorghum plants to avoid or escape the high temperatures in the middle day. Furthermore, various sorghum genotypes demonstrate cooler canopies (escape) or higher canopy temperatures (tolerance), allowing them to either evade or endure excessive tissue temperatures while sustaining higher yields (Prasad et al., 2021). Additionally, sorghum displayed superior salinity-alkalinity tolerance compared to other grass family crops. Some sorghum varieties exhibit strong metal-absorption capabilities, showing potential for phytoremediation in soils contaminated with heavy metals (Mishra et al., 2021; Naseem et al., 2018).

The climate resilience of sorghum is shaped by an intricate interplay of morphological, physiological, and molecular mechanisms. Key traits contributing to this adaptability include a well-developed root system, precise regulation of stomatal activity, and the synthesis and accumulation of osmoprotectants (Enyew et al., 2025). These osmoprotectants, such as proline, glycine betaine, and sugars, are accumulated in plant cells to withstand abiotic stress conditions, including drought, salinity, and extreme temperatures (Enyew et al., 2025).

Sorghum boasts extensive utility across multiple sectors, including food, feed, brewing, and industrial processing. It is classified into four major types based on end-use: grain sorghum, forage sorghum, biomass sorghum, and sweet sorghum (Silva et al., 2022). Grain sorghum, grown primarily for its edible seeds, is a staple food and animal feed in many African and Asian regions. Incorporating sorghum grain into livestock and poultry feed prevents gastrointestinal diseases and improves meat quality. Grain sorghum is essential to China’s liquor production, with most of its output used for brewing, creating a unique liquor style (Meng et al., 2025). Forage and biomass sorghum varieties are valued for their high biomass yield, making them ideal for animal feed through grazing, hay, silage, and bioenergy production (Moore et al., 2020). Sweet sorghum, known for its sugar-rich stalks (crushable for juice extraction and processing), is highly valuable for syrup and biofuel production (Reddy et al., 2005) and has the highest alcohol conversion rate among crops, indicating its potential to impact renewable energy production worldwide (Hu et al., 2022). Sorghum has various industrial applications, including brooms, paper, fibreboard, and natural red pigments. Every part of the plant holds economic value, showcasing its potential for development and commercialization. Thus, sorghum has become essential for food, fodder, energy, and processing industries (Figure 2).

Several key genes and transcription factors have been identified to enhance monocot plant transformation and regeneration efficiency. Notable examples include WUSCHEL (WUS), BABY BOOM (BBM), WOX family members, AGAMOUS-Like 15 (AGL15), and the wound-induced dedifferentiation gene WIN (Nigam et al., 2025). Other promising morphogenic regulators include LEAFY COTYLEDON 1 (LEC1), LEC2, MONOPTEROS (MP), SHOOT MERISTEMLESS (STM), and ISOPENTENYL TRANSFERASE (IPT). Using these factors together can enhance transformation efficiency (Silva et al., 2022). Agrobacterium-mediated transformation of the WUSCHEL 2 (WUS2) gene induces direct somatic embryo formation and regeneration, bypassing genotype-dependent callus formation and shortening the tissue culture cycle (Che et al., 2022; Wang et al., 2023). This method enhances regeneration capacity and transformation efficiency in sorghum. Combined with advanced gene excision systems, such as Cre-LoxP system, it enables high-quality transformation events free of morphogenic genes and selectable markers (Che et al., 2022). Notably, WUS2-assisted transformation bypass genotype-dependent callus formation and significantly shorten the duration of tissue culture achieves a 6.8-fold increase in CRISPR/Cas9-mediated gene knockout efficiency across multiple target sites in various sorghum genotypes (Che et al., 2022). In sorghum, non-integrative BABY BOOM and WUS2 genes overcome the limitations of Agrobacterium transformation based on immature embryo-derived calli of the BTx430 genotype. This method cuts the transformation cycle by 40%, boosts independent transformation events per immature embryo, and offers a breakthrough for sorghum functional genomics and precision breeding (Nelson-Vasilchik et al., 2022). Furthermore, the application of novel promoter combinations to regulate the expression of Wus2 and Bbm promotes the rapid formation of somatic embryos and the regeneration of T₀ plants from seedling-derived early leaf tissues of maize and sorghum after Agrobacterium infection, while this improved leaf transformation method enables Cas9-mediated genome modification in multiple gramineous species including maize and sorghum (Wang et al., 2023b). However, constitutive overexpression of BBM and WUS2 can harm plant development, causing leaf distortion and reduced fertility. Researchers typically remove the BBM and WUS2 expression cassettes before shoot regeneration or use altruistic morphogene-assisted transformation (MAT) to address these issues. These methods increase workload and decrease transformation efficiency due to incomplete removal or insertion of developmental helper genes (Aregawi et al., 2022). It was shown that chimeric sequences of GRF transcription factors and their GIF cofactors significantly improve regeneration efficiency in both monocot and dicot species, expanding transformable varieties and producing fertile transgenic plants. Notably, the GRF4-GIF1 chimera promotes embryogenesis and shoot proliferation in wheat without requiring extra cytokinin supplementation (Debernardi et al., 2020). It was demonstrated that GRF4-GIF1 and GRF5 enhance sorghum transformation efficiency, reducing the process to under two months—an improvement not seen with BBM-WUS systems (Li et al., 2024b). The combination of GRF4-GIF1 and the helper plasmid pVS1-VIR2 achieved the highest transformation efficiency at 38.28%, a 7.71-fold increase, while overcoming growth defects associated with BBM-WUS. Crucially, the CRISPR/Cas9 gene editing tool, developed from the GRF4-GIF1/ternary vector system achieved a 41.36% gene mutation efficiency in sorghum, successfully creating null mutants and offering a stable solution for precision breeding (Li et al., 2024b). Recently, a novel regeneration regulator ZmHSCF1, which promote embryogenic callus formation and proliferation, has been identified. The innovation can be utilized for improving genetic transformation and accelerate crop improvement (Figure 3) (Li et al., 2025). As key regulators of cell fate, morphogenic factors are playing crucial roles in surmounting genotypic restrictions and markedly boosting genetic transformation efficiency in monocots via the induction of somatic embryogenesis and the promotion of recipient cell dedifferentiation.

Compared to conventional breeding, gene overexpression technology offers precise regulation of target gene expression, cross-species introduction of superior genes, and rapid trait improvement, providing powerful technical support for crop genetic enhancement. In addition to the highly efficient genetic transformation methods mentioned above, promoter selection is critical for the precise regulation of gene expression. Constitutive promoters drive gene expression in all tissues, while tissue-specific and inducible promoters selectively regulate expression in specific tissues, developmental stages, or under stress conditions. Constitutive promoters, such as CaMV 35S, are common in dicots but exhibit unstable expression in monocots like sorghum. In contrast, the maize ubi-1 promoter exhibits 10-fold higher expression in monocots, making it a more suitable option (Christensen et al., 1992).

The core function of a tissue-specific promoter is to drive the precise expression of exogenous genes exclusively in specific tissues or organs of plants, while restricting their transcription in non-target tissues. This enables the targeted improvement of transgenic traits and the avoidance of potential adverse effects. Thus, tissue-specific promoters, such as the endosperm-specific α-kafirin and β-kafirin promoters, or the developmentally regulated Sh2 promoter, are increasingly valuable (Liu et al., 2020). The stem-specific A2/LSG promoter drove the expression of vacuole-targeted SUCROSE ISOMERASE (SI), achieving a breakthrough in sugar accumulation. In T₀ transgenic grain sorghum, stems accumulated 50–60% isomaltulose, with total sugar content reaching 1000 mM which is equivalent to an 8-fold increase over controls (118 mM). When elite engineered lines (A5, LSG9) were crossed with sweet sorghum, F₁/F₂ hybrids exhibited >750 mM total sugar, surpassing conventional sweet sorghum (480 mM) and even field-grown sugarcane (600–700 mM) (Liu et al., 2021a). These promoter resources provide essential tools for the precision breeding of sorghum. Future research should focus on optimizing their applications to enable safer, more efficient transgenic crop development.

RNA interference (RNAi) is a powerful gene-silencing mechanism that functions by degrading messenger RNA (mRNA) prior to its translation. In recent years, the RNAi mechanism has emerged as a key gene-silencing tool and been widely applied in the field of plant functional genomics, owing to its conservation across diverse organisms and capacity for sequence-specific gene targeting. For example, Kafirin co-suppression is fundamental to improving sorghum protein digestibility and nutritional value (Elkonin et al., 2021). The nutritional quality of sorghum grain is constrained by a low content of essential amino acids and the protease resistance of its seed storage proteins (kafirins). Targeted RNAi-mediated suppression of specific kafirin subclasses, particularly the γ- and α-types, can effectively improve protein digestibility in sorghum while without compromising critical agronomic traits (Elkonin et al., 2021). RNAi suppression of the opaque2 gene in sorghum embryos lowered kafirin content significantly compared to wild type (Cai et al., 2019). In addition, the transgenic sorghums, which were developed using RNA interference (RNAi) to downregulate genes affecting grain size and protein body structure, had significantly higher crude protein (CP) and higher digestibility, demonstrating the commercial value through genetic engineering (Macelline et al., 2024). Moreover, RNAi technology shows promise for enhancing sorghum biomass. Suppressing 4CL, a key gene in lignin biosynthesis, cut stem lignin by 25% and boosted cellulose and soluble sugar to 36.56% and 59.72%, respectively, significantly improving processing characteristics for forage sorghum (Bhanupriya and Kar, 2025). With the discovery of various classes of regulatory non-coding RNAs (e.g., miRNAs, phasiRNAs, and NAT-siRNAs), the range of RNAi applications has been expanded, which facilitate accurate post-transcriptional and epigenetic modulation (Chaudhary et al., 2024). These achievements provide technical support for sorghum quality improvement. In the future, RNAi will remain a vital component of the functional genomics toolkit and a key complement to gene-editing technologies. We should continue to refine our understanding of RNAi mechanisms, address issues related to targeting effects and long-term stability, improve its precision, expand its scope of application. Fully exploit the enormous potential of RNAi technology in crop genetics.

Gene editing technologies enable precise modifications at the DNA level, including targeted insertions, deletions, or base substitutions (Khalil, 2020). There are three major gene editing platforms: zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs), and the clustered regularly interspaced short palindromic repeats/CRISPR-associated proteins (CRISPR/Cas) system. ZFNs and TALENs were pioneers, but their complexity, high costs, and limited efficiency have led to the more versatile and user-friendly CRISPR/Cas system (Metje-Sprink et al., 2019). The CRISPR/Cas system has been implemented in over 50 plant species. It offers distinct advantages: simplified design protocols, reduced costs, and higher efficiency, making it an indispensable tool for precision breeding in sorghum. Multiple software tools, such as CRISPR-Pv 2, CRISPOR, CRISPRdirect, and CRISPR-PLANT, can design optimal targeting strategies for sorghum genome editing (Liu et al., 2020).

CRISPR/Cas9 editing of the α-kafirin-encoding k1C gene family in sorghum reduced α-kafirin content in seeds, enhancing protein digestibility and lysine levels. The study demonstrated that a single sgRNA effectively edited multiple genes despite mismatches, enabling non-transgenic, nutrient-rich sorghum development (Li et al., 2018). Sorghum in sub-Saharan Africa (SSA) faces severe threats from the parasitic weed Striga hermonthica. LOW GERMINATION STIMULANT 1 (LGS1) is the only known gene locus influencing Striga resistance; loss-of-function alleles (lgs1) reduce Striga germination-stimulating activity. PCR markers detected lgs1 alleles in 6% of 406 sorghum accessions, identifying mutations causing gene loss, including three known deletions (lgs1–1 to lgs1-3) and a novel 50-kbp deletion (lgs1-6) (Adeyanju et al., 2024). CRISPR/Cas9 edited strigolactone (SL) biosynthesis genes (CCD7, CCD8, MAX1) and a DUF gene in the lgs1 region of two sorghum cultivars, achieving 70% transformation and 17.5% editing efficiency. Edited lines exhibited downregulation of SL pathway genes, reduced SL levels in root exudates, delayed Striga infection, and lower infestation rates while maintaining normal growth (Kaniganti et al., 2025). However, suppressing Striga germination by reducing SL biosynthesis leads to abnormal development in sorghum (Chen et al., 2018). Therefore, researchers shifted their focus to limiting SL secretion into the rhizosphere to suppress Striga germination. Shi et al. (2025) identified two ABCG transporter genes crucial for SL secretion to the sorghum rhizosphere. The individual or combined knockout of SbSLT1 and SbSLT2 could significantly impair Striga germination by disrupting SL export. These knockout lines exhibited lower levels of Striga infestation, leading to higher sorghum grain yields and biomass production (Shi et al., 2025). Temporal regulation is crucial for the diverse applications of sorghum cultivated globally. Char et al. (2019) developed a CRISPR/Cas9 system to edit the flowering-time gene SbFT (Sb10G045100) and gibberellin metabolism gene SbGA2ox5 (Sb09G230800) (Char et al., 2019). Persistent Cas9/sgRNA activity induced novel site-specific mutations in progeny, with SbFT mutants showing significant flowering-time variations. Identifying and utilizing male sterility genes is essential for hybrid breeding. CRISPR/Cas9 knockout confirmed that the key male sterility gene, MS8 (Sobic.004G270900), which encodes a conserved bHLH transcription factor, induced male sterility across genetic backgrounds (P<0.01) (Jiang et al., 2021). In China, sorghum brewing value is particularly prominent; “Moutai liquor” and traditional vinegar rely on it. By knocking out SbBADH2 via CRISPR/Cas9, researchers overcame industrial limitations of the aromatic Indian cultivar IS19912, creating new germplasms with aromatic seeds and leaves. Animal trials confirmed improved leaf palatability (Zhang et al., 2022a). Furthermore, Cheng et al. (2024) successfully increased the content of 2-acetyl-1-pyrroline (2-AP) in sweet sorghum via targeted mutagenesis of the SbBADH2 gene; concurrently, they achieved a breakthrough by establishing an Agrobacterium-mediated genetic transformation system and a CRISPR/Cas9-based genome editing system in “Gaoliangzhe” (GZ), an elite sweet sorghum accession, which laid a solid foundation for functional genomic research and biotechnological breeding of sweet sorghum cultivars (Cheng et al., 2024). Recently, Simon et al. (2025) first reported artificially induced apomixis in sorghum, demonstrating that asexual hybrid seeds derived from this approach stably transmit heterosis across generations (Simon et al., 2025). This work is a significant innovation, exemplifying the integrated application of sorghum genetic transformation and genome editing technologies; notably, it holds the potential to capture heterosis and preserve hybrid vigor via sorghum seeds, driving substantial agricultural advancements. Currently, enhanced fertility is a prerequisite for commercial grain production, and further refinements are needed to unlock the full agronomic potential of artificially induced apomictic sorghum in field settings (Simon et al., 2025).

Current genome editing in sorghum mainly focuses on gene knockout, but precision tools like base editing and prime editing show greater promise for accurate mutations. Researchers have created cytidine base editors (CBEs), adenine base editors (ABEs), and dual base editors capable of inducing both substitutions. Examples of these dual systems include SPACE, Target-ACEmax, STEME, A&C-Bemax, AGBE, and ACBE (Chen et al., 2024; Gaudelli et al., 2017; Grünewald et al., 2020; Komor et al., 2017; Li et al., 2020; Liang et al., 2022; Sakata et al., 2020; Zhang et al., 2024d, Zhang et al., 2020). Beyond base editing, promoter editing can also establish effective quantitative trait variations. Zhou et al. (2023) developed a computational model that assigns values to different promoter regions and created a CRISPR-Cas12a-based promoter editing system to guide promoter editing experiments for fine-tuning gene expression and generating desired quantitative traits (Zhou et al., 2023). Recently, a transgene-free genome editing system was established in sorghum through targeting PHYTOENE DESATURASE (PDS) gene to generate a visible phenotype mutation albino (Zhang et al., 2025b). Field-grown transgenic sorghum exhibits a high risk of unintended hybridization with closely related weedy relatives (e.g., Sorghum bicolor subsp. drummondii, commonly known as Sudan grass), which may result in the spread of transgenes and subsequent transgenic contamination in agricultural ecosystems (Wolfenbarger and Phifer, 2000). In sharp contrast, precision genome-edited sorghum lines can be generated without the integration of exogenous transgenes (He et al., 2022; Li et al., 2024a; Zhang et al., 2025b), thus circumventing the biosafety concerns associated with transgenic flow and hybridization. The production of transgene-free, site-specific edited sorghum germplasm therefore represents a critical strategy for mitigating the risks of transgenic contamination while retaining the agronomic benefits conferred by targeted genetic modifications. Given its robust editing efficiency, this system holds significant promise for future applications in sorghum improvement.

Genomic analysis offers a fundamental blueprint of genomic information, serving as the foundation for molecular breeding. Genomic analysis of cereals enables the screening of wild and domesticated germplasms to identify novel sources of desirable traits. The first sorghum reference genome (from grain sorghum BTx623) was sequenced in 2009, with approximately 98% of genes anchored to chromosomes (Table 1) (Paterson et al., 2009). An improved version (BTx623 v3.1.1) was released in 2018 (McCormick et al., 2018). Subsequently, a chromosome-scale de novo assembly of the repeat-rich Tx430 genome was achieved by integrating Oxford Nanopore MinION sequencing data with Bionano Genomics Direct Label and Stain (DLS) optical maps in 2018 (Deschamps et al., 2018). Additionally, a high-quality reference genome for the sweet sorghum cultivar “Rio” was completed using Pacific Biosciences long-read sequencing technology in 2019 (Table 1) (Cooper et al., 2019). These genomic resources have been widely utilized to identify various genetic variations, including single-nucleotide polymorphisms (SNPs), insertions/deletions (InDels), structural variations (SVs), presence/absence variations (PAVs), and copy number variations (CNVs).

The complete telomere-to-telomere (T2T) assemblies enhance our understanding of genome structure, biology, and agricultural applications. Advancements in sequencing technologies, like PacBio HiFi (base accuracy >99.9%), ultra-long Oxford Nanopore Technology (UL-ONT; >100 kb), and Hi-C, have made T2T genome assemblies possible. Researchers assembled the BTx623 genome using Hifiasm with ultra-long ONT data (~82.2x coverage) and PacBio HiFi data (110.7x), then polished and corrected contigs using Hi-C (127.2x) and Illumina sequencing (65.4x), achieving a complete T2T assembly of all 10 chromosomes, including telomeres and centromeres (BTx623-T2T) (Table 1) (Deng et al., 2024). The BTx623-T2T genome covers 719 Mb, incorporating 43.6 Mb of new sequences compared to the previously published BTx623-v3.1, mainly in complex regions such as centromeres, telomeres, and other repetitive areas. Additionally, BTx623-T2T closed 3,913 gaps and corrected 1,131 misassemblies in BTx623-v3.1. Using transcriptome data from 76 multi-tissue/stage samples, 3,565 new protein-coding genes were annotated. The BTx623-T2T reference genome significantly improved the utilization of Illumina reads and alignment accuracy compared to BTx623-v3.1. A GWAS study on stem water content in 202 sorghum accessions found that using the BTx623-T2T reference genome identified all significant loci from BTx623-v3.1 and revealed three additional loci, highlighting its effectiveness for candidate gene discovery in agronomic traits (Table 1) (Deng et al., 2024).

The Hongyingzi (HYZ) variety, developed in 2008, has high grain tannin content and other desirable traits for distilled liquor (Baijiu) production, accounting for over one-third of China’s sorghum cultivation (Zhang et al., 2022b). Ding et al. performed a high-accuracy T2T assembly of the HYZ genome using UL-ONT, PacBio HiFi, and Hi-C. The complete de novo HYZ genome assembly fills gaps in tannin synthesis pathways and identifies genetic targets for Baijiu-focused breeding. Subsequently, Bao et al. released the Sorghum T2T Genome Database (http://sorghum.org.cn/), featuring high-quality T2T assemblies of two Chinese Baijiu landraces: Hongyingzi (used for Maotai, China’s most famous Baijiu) and Huandiaonuo (used for Fen-flavor Baijiu) (Bao et al., 2024). T2T resources provide insights into dark genomic regions (e.g., centromeres/telomeres), the highest-resolution genetic map for Baijiu-trait breeding, and a model for T2T assemblies in other plants. Comparative genomic analysis reconstructed a 65-gene metabolic pathway for tannin synthesis (including eight transcription factors, three transporters, and 45 structural genes), providing a genetic reference for Baijiu-oriented breeding (Ding et al., 2024). Three regulatory genes (Yellow seed 1, TANNIN 1, AND TANNIN 2) involved in tannin biosynthesis have been identified (Zhang et al., 2024b, Zhang et al., 2023b).

Recently, Chen et al. (2025) assembled a gapless, T2T reference genome (729.5 Mb) for sorghum variety E048. Comparative genomic analysis identified a 2.9 Mb E048-specific region with 187 genes related to signal transduction, immune response, and metabolic regulation (Chen et al., 2025). This offers insights into superior agronomic traits like disease resistance, stress tolerance, lodging resistance, and high sugar content. They developed an EMS mutant library comprising 13,226 M₁ plants (covering 97.54% of genes), established efficient MutMap/MapMap+ gene mapping methods, and optimized an Agrobacterium-mediated transformation system. These resources provide a “genome-mutant-transformation” pipeline for functional genomics and molecular breeding in sorghum. All data have been integrated into the Sorghum Genome and Mutant Database (SGMD; https://sorghum.genetics.ac.cn/SGMD), creating a comprehensive research platform that links genome analysis and breeding applications (Table 1) (Chen et al., 2025).

A single reference genome cannot capture the full genetic diversity of a species, limiting exploration of genetic variations. Additionally, restricted diversity from recombination in elite breeding populations may fail to address environmental challenges. The pangenome of a species consists of all its genes and is essential for understanding variation within it. It has three components: (1) core genome (genes shared by all individuals), (2) dispensable genome (genes present in some individuals), and (3) strain-specific genes (unique to single strains) (Tao et al., 2021a). A representative sorghum pan-genome had been published in 2021 (Tao et al., 2021a). Tao et al. (2021a) constructed the first broadly representative sorghum pan-genome using 13 varieties, including S. propinquum, wild sorghum, and cultivated sorghum. They employed multiple omics technologies, including second-generation sequencing, third-generation sequencing, Hi-C, and transcriptomics, achieving a maximum contig N50 of 3.48 Mb for genome assembly. This sorghum pan-genome size is 954.8 Mb, 30% larger than the reference genome (BTx623, 732.2 Mb), with core sequences accounting for 62% and extensive presence-absence variations (PAVs) across genomes. Combining pan-genome data, they performed GWAS on grain color and identified a 3,216 bp PAV in the Yellow seed1 gene. SbRC, homologous to the rice grain color gene Rc, showed a 416 bp PAV in the pan-genome.

Ruperao et al. (2021) developed a more comprehensive sorghum pan-genome using reference genomes and 354 genetically diverse sorghum accessions representing different races. They identified 35,719 genes, with a core genome of 16,821 and an average of 32,795 per cultivar. Notably, 53% of genes exhibited presence-absence variation, with variable genes enriched for environmental responsiveness and capable of classifying accessions by race. Association analysis using over two million SNPs from the pan-genome identified 398 significant SNPs linked to agronomic traits. The expression analysis of drought-responsive genes revealed 1,788 are functionally essential, including 79 absent from the BTx623 genome, providing valuable genomic resources linking genetic diversity to adaptive traits, particularly drought response (Ruperao et al., 2021). Marla et al. (2025) recently analyzed a pan-genome derived from 1,661 sorghum germplasm resources and identified seven loss-of-function variants in the DW3 gene. A 137 bp deletion Dw3 allele from the variety Segaolane showed excellent properties for suppressing plant height reversion mutations (Marla et al., 2025). This study provides an innovative solution to the long-standing problem of height reversion mutations in U.S. grain sorghum breeding. The sorghum pan-genome will revolutionize research by integrating GWAS of key traits with structural variation analysis, advancing our understanding of sorghum functional genomics and breeding (Table 1) (Poosapati et al., 2022).

Multi-omics involves studying organisms at multiple molecular levels, such as genomics, transcriptomics, proteomics, and metabolomics. SorghumFDB (http://structuralbiology.cau.edu.cn/sorghum/) is a functional genomics data mining platform that provides comprehensive functional annotations of genes (Tian et al., 2016). Liu et al (2021b) integrated whole-genome SNP and INDEL variation data from 289 sorghum accessions based on the BTx623 (v3.1) reference genome, establishing the multi-omics platform SorGSD (https://ngdc.cncb.ac.cn/sorgsd/). It integrates phenotypic traits and panicle morphology images for genotype-phenotype co-analysis. It uses tools like ID conversion, gene alignment, and genome browsers, enabling efficient multi-omics data mining for sorghum functional genomics and molecular breeding (Liu et al., 2021b). Bao et al. (2024) established the first brewing sorghum genome database (http://sorghum.org.cn/) featuring genome browsing, sequence alignment, chromosomal collinearity analysis, and data download (Table 1) (Bao et al., 2024). Chen et al. (2025) developed the Sorghum Genomics and Mutant Database (SGMD, https://sorghum.genetics.ac.cn/SGMD) by integrating genomic data, a gene expression atlas, and mutant variations, serving as a “super toolkit” for sorghum functional gene research (Table 1) (Chen et al., 2025).

Breakthroughs have been made in understanding sorghum stress resistance mechanisms through multi-omics approaches. Jiao et al. (2023) used multi-omics analysis to identify 2,683 differentially expressed genes and 160 metabolites, offering insights into crop cadmium resistance mechanisms (Jiao et al., 2023). Li et al. (2024a) revealed that drought and salt stress responses involved whole-genome duplication and conserved domains in sorghum transcription factors, while transcriptomics identified 45 key genes. RNA and degradome analyses revealed miR5072 and its target gene (Sobic.001G449600) related to drought resistance, while WGCNA further identified drought-responsive genes. Ultimately, 15 candidate genes were found, including two TFs: HD-ZIP family Sobic.004G300300 and bZIP family Sobic.003G244100 (Li et al., 2024c). Liu et al. (2024) integrated genomics and transcriptomics with phenotypic and physiological analyses to elucidate nitrogen use efficiency (NUE) mechanisms in sorghum. Co-expression network analysis identified key genes like nitrogen transporter Sobic.003G371000.v3.2leaf (NPF5.10) and transcription factor Sobic.002G202800.v3.2leaf (WRKY) that enhance NUE by regulating nitrogen uptake (NUpE) and utilization efficiency (NUtE) under low-nitrogen stress, presenting vital targets for developing N-efficient sorghum varieties (Liu et al., 2024). Seitz et al. (2024) used metabolomics and metatranscriptomics to analyze how root exudates from four crops (sorghum, hairy vetch, rapeseed, and rye) regulate soil microbiome functions. They created the first genomic database on crop-microbiome interactions, which helps to understand soil biogeochemical processes (Seitz et al., 2024). Mishra et al. (2021) studied micronutrient responses in sorghum under iron (Fe) and zinc (Zn) deficiency/excess by integrating transcriptomics and ionomics. RNA-seq revealed transcriptional regulation in roots and leaves during stress, with Fe deficiency and Zn excess causing notable phenotypic and gene expression changes. Gene regulatory network (GRN) analysis identified hub genes (SbFIT, SbPYE, SbBTS) in roots that regulate Fe/Zn uptake, while leaf homologs primarily influence chloroplast function, photosynthesis, and oxidative stress. Xing et al. (2025) used multi-omics analysis to show how spermidine (Spd) improves vigor in aged sorghum seeds via antioxidant networks. Integrated analysis of transcriptomics, proteomics, and metabolomics shows that Spd enhances antioxidant enzyme activity and metabolite accumulation, clearing ROS and reversing seed aging effects (Xing et al., 2025).

Shi et al. (2024) developed HEMU (The HEMU Andropogoneae Database), the first integrated multi-omics analysis platform for Andropogoneae grasses, incorporating nearly 5,000 multi-omics datasets across 20 species (including maize, sorghum, and sugarcane). HEMU covers 75 genomes, 4,718 RNA-seq datasets (1,527 maize, 1,428 sorghum), 90 ChIP-seq epigenomic datasets (14 tissues, 10 histone modifications), and 37 ATAC-seq datasets (13 tissues). Its innovative six-toolkit system (genomics, transcriptomics, epigenomics, gene families, transposable elements, and integrated analysis) enables multi-level, one-stop analysis from DNA sequences to epigenetic modifications, providing powerful multi-omics support for functional gene discovery and molecular breeding in Andropogoneae crops (Zhu et al., 2024). Ko and Brandizzi (2025) developed NEEDLE (https://github.com/DaeKwan-Ko/needle) to tackle analytical challenges in non-model crops. This tool integrates dynamic transcriptome data to construct co-expression network modules and identify key transcriptional regulators, successfully analyzing cellulose synthase-like F6 (CSLF6) in sorghum and Brachypodium (Table 1) (Ko and Brandizzi, 2025).

Recent innovations in molecular biology and genomics have greatly improved our understanding of the genetic regulatory networks governing yield-related traits in sorghum. Plant height, which influences lodging resistance and biomass, has garnered widespread attention in breeding. Quinby and Karper identified four loci (Dw1-Dw4) that regulate plant height by modifying internode length (Karper and Quinby, 1947). However, the unclear allelic composition of Dw1-Dw4 genes in primary breeding materials has increased plant height in the identical height type progeny, leading to suboptimal morphological traits. Wang et al. (2024) revealed that China’s predominant sterile lines mainly exhibit the “triple-dwarf” type (Dw1-Dw2-dw3-dw4) from Kafir and its improved lines, while restorer lines are primarily composed of the improved “double-dwarf” type (Dw1-Dw2-dw3-dw4) from the Kaoliang/Caudatum subspecies, along with some Kafir-derived “triple-dwarf” types. Notably, the dw3 allele was predominant in the tested materials, whereas dw1 occurred less frequently in the restorer lines. Importantly, the dw2 allele, which significantly influences plant architecture, was completely absent in key restorer materials. These findings highlight the need for precise genotyping of Dw1 and Dw2 alleles to enable differentiated breeding strategies, offering theoretical and technical support for marker-assisted dwarf breeding sorghum (Table 2) (Wang et al., 2024). Complementing the Dw1-Dw4 system, several other genes regulating plant height have been identified in sorghum, which controlplant height by regulating internode elongation and cell proliferation. A 740-bp transposable element insertion in qHT7.1 (encoding a MYB transcription factor) intron leads to aberrant splicing and premature termination, resulting in a dwarf phenotype in sorghum.

Grain size variation is a major determinant of yield and quality in cereal crops. It is governed by both the plant’s genetic potential and the availability of assimilates allocated for grain filling. Tao et al. (2021a) found that five grain-size-related parameters exhibited high heritability, and artificially reducing grain number led to increased grain weight. The GWAS analysis identified 94 QTLs, with SbDEP1 confirmed to balance grain number per panicle and grain weight by regulating primary branch number. It provided insights on “source-sink” relationships, showing that grain size is influenced by genetic potential and assimilate partitioning, highlighting key targets for improving yield components in cereals (Table 2) (Tao et al., 2021b).

Grain number per panicle is another critical yield determinant. Jiao et al. (2018) obtained a multi-seeded sorghum mutant (msd1) through EMS mutagenesis and discovered that MSD1, a TCP-family transcription factor, influences panicle development via the jasmonic acid (JA) pathway. The mutant exhibited a 50% reduction in JA content in young panicles, and exogenous JA application restored the phenotype. This finding establishes the first link between JA signaling and panicle architecture (Jiao et al., 2018). The DG1 locus promotes lower floret development by modulating histone modifications, leading to a double-grain trait. These studies elucidate molecular mechanisms of inflorescence development and highlight the unique value of multi-grain sorghum in brewing (Table 2) (Zhang et al., 2025a).

Seed shattering is a trait that directly impacts harvesting efficiency. Lin et al. (2012) discovered that seed shattering is controlled by a single gene, SHATTERING1 (SH1), which encodes a YABBY transcription factor in sorghum. Domesticated sorghum carries three distinct mutations at the SH1 locus, and variations in the promoter and intronic regulatory regions result in low expression levels. A 2.2-kb deletion leads to a truncated transcript lacking exons 2 and 3. A GT-to-GG splice-site mutation in intron 4 causes the exclusion of exon 4. SH1 underwent parallel selection during the domestication of sorghum, rice, and maize. Notably, the 2.2-kb deletion mutation in 80% of cultivated varieties significantly reduces harvest losses, providing molecular evidence for understanding crop domestication (Table 2) (Lin et al., 2012).

Threshing efficiency in Poaceae crops relates closely to seed hull enclosure. Cereal crops like sorghum, rice, and wheat typically have seeds enclosed by glumes, with hull loss marking a significant event in panicle domestication. Sorghum exhibits rich phenotypic variation in hull enclosure. GWAS analysis identified GC1 as a negative regulator of hull enclosure, encoding an atypical Gγ subunit. GC1 modulates hull cell proliferation; overexpression reduces hull enclosure, while knockout enhances it. GC1 interacts with phosphatase pPLAII-1, promoting its degradation to regulate glume development. gc1 allelic variants are present in 40% of sorghum germplasms, indicating strong artificial selection for this beneficial characteristic (Table 2) (Xie et al., 2022).

Optimizing plant architecture is a critical pathway for achieving yield gains. Studies in sorghum (Sorghum bicolor L. Moench) show that high-density planting of erect hybrid varieties mainly drives increased yields. Through CRISPR/Cas9, researchers successfully generated monoallelic and biallelic mutants of the LG1 gene in sorghum. Monoallelic mutants exhibited enhanced leaf erectness, while biallelic mutants lacked ligule structures and showed further reduced leaf angles (Brant et al., 2021). However, high-density planting triggers shade avoidance responses (SAR) that optimize light capture but compromise plant vigor and ultimately limit yield potential. The mechanism behind this phenomenon is that phyB1/B2 serves as the primary photoreceptor, detecting changes in the ratio of red (R) to far-red (FR) light and coordinating plant responses through the LG1-HB53 regulatory module (Shi et al., 2024). Plant architecture influences light-use efficiency in sorghum and closely correlates with panicle development. Panicle neck elongation is a critical yield factor. The sheathed panicle-I (shp-I) mutant showed shortened neck internodes due to reduced parenchyma cell size. A single recessive gene, SbiHYZ.10G230700, controls this trait, encoding a BTB/POZ and MATH domain protein. Intriguingly, the mutant also exhibited reduced auxin levels and elevated brassinosteroids, suggesting synergistic hormonal regulation of panicle neck development (Table 2) (Ao et al., 2025).

Tillering directly affects plant structure and yield formation. A dual-pathway regulatory model was established for Sorghum tillering (Kebrom et al., 2010, Kebrom et al., 2006). Kebrom et al. (2006) first revealed the phyB-SbTB1 module that controls sorghum tillering, where the active form of phytochrome B (phyB, Pfr) inhibits SbTB1 to promote tiller bud, while inactive phyB increases SbTB1 to suppress tillering (Kebrom et al., 2006). Later, a separate pathway was identified where defoliation regulates tillering through the SbMAX2 gene (an Arabidopsis MAX2 homolog) mediated strigolactone signaling pathway (Kebrom et al., 2010). Chen et al. (2018) added NAB1 (encoding CCD7 enzyme) as a crucial regulator; its mutant increases tillering due to disrupted strigolactone biosynthesis and enhanced auxin polar transport (Zhang et al., 2019). Regarding transcriptional regulation, comparative analysis revealed the conserved function in grasses of the TIN1 gene, which encodes a C2H2 zinc finger transcription factor to regulate tillering by suppressing gt1 and Laba1/An-2 expression and interacting with TOPLESS proteins. A major QTL for tillering number was identified in the TIN1 region in sorghum (Zhang et al., 2019). Subsequent QTL mapping identified regions with differentially expressed genes (DEGs), including potential regulators like DRM1 and WUSCHEL. With advances in high-throughput sequencing technologies, researchers have begun to unravel the genetic basis of tillering traits at the whole-genome level. GWAS (Wang et al., 2022; Wondimu et al., 2023) using SNP markers identified multiple stable QTNs for tiller number (Wang et al., 2022; Wondimu et al., 2023). Upadhyaya et al. (2024) evaluated a mini-core sorghum collection and found a consistently detected tillering (TL) locus on chromosome 1, containing Sobic.001G152700 (encoding a DUF1618 protein). This gene is the only gene horizontally transferred from Sorghum to the parasitic weed Striga hermonthica, potentially related to environmental adaptation (Upadhyaya et al., 2024).

Panicle morphology in sorghum influences its grain yield and resistance to pests and diseases. Panicle morphology traits include length, rachis node number, primary branch number, maximum primary branch length, and compactness. GWAS identified that 71 QTLs were distributed across 41 genomic regions on 9 chromosomes using a sorghum population adapted to diverse environments in China (Zou et al., 2024). Two domestication-related genes (Sobic.003G052700 and Sobic.006G247700) were located within two major QTL regions (QTL3.4721839 and QTL6.58709500) detected across multiple environments. The allelic variations of these genes exhibited a geographical pattern, suggesting that sorghum breeders in southern and northern China have selected for different panicle morphology traits. Southern sorghum varieties have long, loose panicles, adapting to hot, humid climates, while northern varieties exhibit short, compact panicles, enabling higher planting density and greater grain yield in arid regions (Zou et al., 2024). This work offers new breeding strategies and resources for developing sorghum suited to local conditions.

Plant height and branching characteristics jointly determine canopy structure and light-use efficiency. The genetic regulation of energy metabolism pathways can coordinate these architectural traits with reproductive growth. Upadhyaya et al. (2022) conducted a GWAS analysis using a sorghum mini-core collection and identified multiple QTLs for days to flowering, plant height, biomass, and sugar content. Notably, overexpression of Sobic.006G061100 (SbSNF4-2, encoding the γ-subunit of the AMPK/SNF1/SnRK1 complex) in both sorghum and sugarcane significantly increased biomass and plant height while delaying flowering and enhancing sugar content. It was revealed how energy-sensing pathways integrate plant development with carbon partitioning, providing key targets for achieving an ideal plant type characterized by “high biomass-high sugar-moderately late flowering” (Upadhyaya et al., 2022).

Importantly, yield improvement relies not only on aboveground architectural optimization but also on belowground organ functionality. Conventional studies suggest a trade-off between rhizomes (as carbon storage organs in perennial crops) and grain yield. However, Zheng et al. found that rhizome biomass exhibits high heritability (0.723) and strong positive correlations with total belowground biomass (r1 = 0.95; r2 = 0.97). A positive correlation was found between rhizome biomass and grain yield, potentially mediated by rhizome-enhanced tillering effects. Through bulked segregant analysis (BSA), researchers mapped 20 SSR markers linked to rhizome traits in 8 genomic regions, including 5 novel loci, and selected elite lines with high rhizome biomass, biomass yield, and grain yield (Zheng et al., 2025). This discovery challenges traditional paradigms, proving that we can improve sorghum’s carbon sequestration and agronomic yield through genetic breeding.

The nutritional value of sorghum is limited by three factors: lack of essential amino acids (notably lysine), low protein digestibility, and insufficient sugar and oil in traditional varieties (Duodu et al., 2003). Nutritional shortcomings have caused widespread “hidden hunger” in areas where sorghum is a staple—malnutrition marked by adequate energy intake but micronutrient deficiencies. Enhancing sorghum’s nutrition via molecular breeding and metabolic engineering has become a new frontier in global agricultural research.