Ramasamy Perumal

Sorghum ( Sorghum bicolor (L.) Moench) is a diploid crop that is primarily grown in regions with high temperatures and limited irrigation. Breeding efforts often focus on stay‐green, a trait associated with green leaf area after the crop experiences post‐flowering drought. When present, this trait is assessed using subjective visual categorical ratings. The objective of this study was to assess stay‐green across nine environments and to develop a machine learning model that can estimate stay‐green using vegetative indices collected by an unoccupied aerial system. Hybrids were derived from a partial factorial design that included material from three public and one private breeding program. Results showed that each program had heritable variation for this trait and the benefit of sharing material to develop top‐performing stay‐green hybrids varied among each breeding program. Five machine learning models were built to estimate stay‐green; they produced accuracies up to 92% (gradient boosting model) in testing and training. The validation dataset, which contained data from four unseen environments, produced average accuracies up to 66% (naïve Bayes model). The naïve Bayes model stay‐green estimates also resulted in a heritability 16% higher than that of the field‐collected ratings. Overall, the naïve Bayes and gradient boosting models performed best, demonstrating potential for more efficient and less biased stay‐green rating collection in grain sorghum.

Early planting in the U.S. Midwest has the potential to improve sorghum productivity through effective use of residual soil moisture and avoidance of late-season heat and drought stress. However, exposure to sub-optimal temperatures early in the season often reduce seedling establishment and yield. Therefore, identifying sorghum hybrids with early-stage chilling tolerance and stable grain yield is critical. Herein, 215 grain sorghum hybrids were generated in a partial factorial mating design of 41 females and 16 males from four breeding programs (Kansas State University (KSU), Texas A&M System (TAM), USDA-ARS (ARS), and S&W Seed Company (SW)). Hybrids were evaluated in controlled environments (control vs. chilling stress) and in field conditions across two locations using early (late April) and regular (late May) planting dates. Several seedling traits (e.g., emergence percentage, emergence speed index, and seedling biomass) assessed early-stage chilling tolerance in controlled environments and field conditions, while days to flowering, plant height, and grain yield were also measured in the field trials. Mixed model analysis revealed significant effects of planting window and hybrid, with early planting generally reducing seedling performance and grain yield. Hybrid performance rankings varied across planting dates and environments. Variance partitioning showed that both additive (GCA, general combining ability) and non-additive (SCA, specific combining ability) genetic effects contributed to seedling traits, while grain yield in the early planting was influenced by SCA and interaction effects. A two-step selection pipeline combining multi-trait stability index and GGE (genotype and genotype by environment) biplot analysis identified chilling-tolerant, high-yielding, and stable hybrids. ARS82A/KSU69R and KSU52A/KSU64R were consistently identified across selection pipelines, along with other promising hybrids (KSU50A/KSU61R, S-W11A/KSU72R, and TAM33A/KSU62R). Among the selected chilling tolerant hybrids, most were derived from cross-breeding program combinations, indicating the value of integrating diverse germplasm. Moreover, some chilling tolerant hybrids were identified from crosses in which neither parent had been specifically selected for chilling tolerance, suggesting that germplasm exchange can generate favorable hybrid combinations and enhance genetic gain.

Early‐season chilling limits sorghum establishment in temperate environments, yet the metabolic basis of this sensitivity remains poorly understood. We combined temporal spectral reflectance with untargeted metabolomics to characterize species and genotype‐level responses to early planting conditions. Normalized Difference Vegetative Index (NDVI) and Normalized Difference Red‐Edge (NDRE) index revealed consistently faster seedling development and greater early‐season greenness in maize compared with sorghum, with chilling‐tolerant sorghum genotypes exhibiting intermediate reflectance patterns. Principal component analysis of metabolite profiles separated maize and sorghum along a dominant species axis, while planting date effects and genotype‐specific metabolic stability were captured along secondary axes. Untargeted metabolomics identified 144 differentially accumulated metabolites (DAMs) between maize and sorghum and 56 DAMs between early‐ and optimum‐planted sorghum. Early planting induced marked metabolic reorganization in sorghum, characterized by the accumulation of phenylpropanoid intermediates (chlorogenic acid, 4‐hydroxybenzoic acid), osmoprotective sugars and polyols (galactinol, mannitol), amino acid derivatives (pipecolic acid, O‐acetylserine), and aromatic amino acids, accompanied by depletion of central carbon metabolites including sucrose, fructose, sorbitol, and 3‐phosphoglycerate. Pathway enrichment analysis indicated chilling‐induced suppression of starch and sucrose metabolism, fructose–mannose pathways, and amino sugar metabolism, alongside activation of phenylpropanoid, terpenoid, aromatic amino acid, glucosinolate, and cyanoamino acid biosynthesis in sorghum compared with maize. The results of the current study demonstrate that early‐season chilling constrains carbon turnover in sorghum and promotes diversion of metabolic flux toward antioxidative and osmoprotective pathways. The spectral metabolomic framework presented here identifies biochemical signatures associated with chilling sensitivity and provides targets for improving early‐season vigor in sorghum.

Fusarium stalk rot in sorghum is caused by Fusarium thapsinum and affects sorghum production globally. In this study, we investigated genome-wide transcriptional responses associated with resistance to sorghum stalk rot. Transcriptomes of two sorghum genotypes, a resistant (SC599) and a susceptible (BTx3042) genotype, were analyzed at 24 and 48 h post-inoculation (hpi) with F. thapsinum at flowering. A total of 24 RNA-seq libraries were generated. Differentially expressed genes (DEGs) ranged from 360 to 6,178 (FDR < 0.05), with earlier and stronger transcriptional responses observed in the resistant genotype. Key defense-related genes, including nucleotide-binding site leucine-rich repeat (Sobic.005G226400), pathogenesis-related proteins (Sobic.001G401200), disease resistance-response genes (Sobic.009G021300), wall-associated kinase (Sobic.008G170800 and Sobic.002G249600), and flavonoid synthase (Sobic.004G050200) genes, were upregulated in SC599 at 24 hpi, whereas expression of these genes was lower or delayed in BTx3042. Gene ontology and KEGG analyses revealed enrichment of defense responses, flavonoid biosynthesis, and signal transduction at 24 hpi in the resistant genotype, whereas the susceptible genotype showed delayed activation of stress-related pathways. These results highlight the importance of early and coordinated activation of defense responses and provide insight into the temporal regulation of resistance mechanisms in sorghum against F. thapsinum and may provide a framework for future studies to understand the genes and pathways underlying the molecular mechanisms of Fusarium stalk rot resistance in sorghum.

Drought and heat stress are major challenges to crop productivity, especially in semi-arid regions with limited water and erratic rainfall. Pearl millet, the sixth most widely cultivated cereal for food, forage, and feed, has emerged as a promising climate-resilient crop to address these challenges. However, the genomic regions associated with its drought-tolerance remain largely unknown. This study utilizes genome-wide association analysis (GWAS) to identify drought related quantitative trait nucleotides (QTNs) in 187 pearl millet germplasms, genotyped using genotyping-by-sequencing (GBS), yielding 35,071 high-quality single nucleotide polymorphic (SNP) markers. Phenotypic variation was observed between irrigated and rainfed treatments, and the stress tolerance index (STI) derived from these data was used for GWAS. Ninety-five QTNs were mapped across all seven chromosomes using four GWAS models (MLMM, FarmCPU, Blink, and 3VmrMLM). These QTNs co-localized in close proximity with 86 candidate genes, including PMF1G04719 and PMF2G07960 (rubisco binding protein), PMF1G07862 (actin-7). These were involved in drought response mechanisms, such as reactive oxygen species (ROS) scavenging, and abscisic acid signaling pathway. Candidate gene haplotype analysis identified 23 superior haplotypes with higher drought tolerance traits, offering promising targets for further investigation and breeding efforts. These findings serve as valuable genetic and genomic resources for improving pearl millet for climate resilience and production in semi-arid regions.

Pearl millet (Pennisetum glaucum L. R. Br.), a climate-resilient crop with rich nutritional clues, that can be a promising source for dietary diversification. This study was aimed at systematic measurement of genetic variation for key nutritional traits and therapeutic clues in a diverse panel of 232 pearl millet genotypes and identifying genetic loci linked to the accumulation of key nutritional traits through metabolome based genome-wide association study (mGWAS) and identification of putative candidate genes through haplotype analysis. GWAS analysis using a total of 30,488 high-quality single nucleotide polymorphisms (SNPs) obtained through genotyping by sequencing with a minor allele frequency (MAF > 0.05) identified 124 significant quantitative trait nucleotides (QTNs) linked to 13 different nutritional traits. putative candidate genes linked to lipid/antioxidant traits were found to be involved in fatty acid elongation, desaturation, phytosterol metabolism and phenolic acid biosynthesis. Haplotype analysis detected novel allelic variantions influencing accumulation of secondary metabolites. Pearl millet genotypes viz., ARCH75R, ARCH12B and PI197417 exhibited superior lipid and antioxidant profiles, highlighting their potential for utility in breeding for biofortification in pearl millet. These findings provide novel insights into the molecular mechanisms regulating accumulation of bioactive compounds in pearl millet and paved way for developing high yielding and nutrient-dense pearl millet, genotypes through molecular breeding.

Pearl millet, recognized for its resilience to climate challenges, thrives in drought-prone and poor soil quality areas. This study aims to evaluate the impact of drought stress on six agronomic traits in 188 pearl millet accessions and identify stable and drought-tolerant accessions that can be developed into forage parental lines. Field experiments conducted in summer 2023 and 2024 under irrigated (control) and rainfed (stress) treatments at Hays, Kansas, revealed significant differences among the accessions, treatments, and year based on an aligned rank transform test for all the traits evaluated. The rank summation index selected 30 drought-tolerant germplasms per maturity group from the whole collection. Principal component analysis based on GBS SNP data indicated a higher level of genetic variation within the selected and unselected germplasms. By utilizing BLUP-based GGE biplot, MTSI, MGIDI, and FAI-BLUP stability models, we identified three early (E11—PI197115, E5—PI197368, and E12—PI197327), four medium (M19—PI197465, M5—PI197436, M1—PI197326, and M29—PI197069), and two late (L5—PI197291 and L19—PI197278) maturity accessions with broad adaptability across environments. Chlorophyll index, stay-green, and biomass traits served as effective indicators for drought tolerance, showing moderate to high heritability across maturity groups and stable performance across years. These results are crucial for developing drought-tolerant parental lines and hybrids with enhanced forage yield, addressing livestock feed challenges.

Sorghum [ Sorghum bicolor (L.) Moench] is the fifth most important cereal grain crop worldwide and is used as both feed and food grain. While grain composition and quality are important, they have traditionally been a lower priority relative to grain yield. If methods to predict composition and quality are available, this could be added to the selection criteria with minimal addition of time or cost. Herein, the genetic inheritance of sorghum grain quality traits was assessed in hybrids obtained by crossing 10 elite inbreds from Texas A&M and Kansas State University following factorial mating designs. Grain samples from these 100 hybrids were collected from 10 evaluation environments and then analyzed for starch, protein, fat, and fiber using near‐infrared spectroscopy. In addition, grain samples were characterized for three physical factors: kernel hardness index (KHI), kernel diameter (KD), and kernel weight (KW). Environmental effects were a major source of variation for starch (33.9%), fat (53.5%), and fiber (53.9%), whereas genetic effects were prominent for protein (29.9%), KHI (59.7%), KD (56.9%), and KW (43.8%). Starch, fiber, KHI, and KD predictions were more accurate (0.38–0.74) than those for protein, fat, and KW (0.26–0.70). Finally, multi‐trait genomic selection models that included grain yield and days to anthesis improved prediction accuracies up to 18% for grain quality traits over single‐trait models. In conclusion, these genomic selection models have the potential to effectively and concurrently select for grain composition and quality factors in sorghum.