Genome-wide association analyses reveal genetic loci associated with cold responses in Miscanthus under non-acclimated and cold-acclimated conditions

Miscanthus species are C4 perennial bioenergy crops with broad environmental adaptability, yet episodic freezing events can constrain establishment and productivity. We evaluated cold-response variation in a 144-accession panel comprising 73 M. sacchariflorus and 71 M. lutarioriparius accessions under non-acclimated and cold-acclimated conditions. LT50 was recalculated from the original mean relative electrical conductivity (REC) data using y = 100/[1 + a exp(bx)] and LT50 = -ln(a)/b; fits were retained when R² >= 0.80 and LT50 fell within the experimental range (-16 to 0 °C). Additional genotype quality control removed two accessions with sample missingness > 15% and variants with extreme observed heterozygosity ( > = 90%), yielding a final panel of 142 accessions and 8,043,797 SNPs. Primary GWAS used GEMMA with PC1-PC3 and genomic kinship; trait-specific tested marker counts ranged from 7,470,441 to 7,963,728, while a conservative fixed Bonferroni threshold based on the full panel-level marker count was applied ( P < 6.22 × 10 − 9 ). Three of 17 traits - NonACC_0, NonACC_m4, and NonACC_m8 - showed genome-wide significant associations, comprising 26 significant SNPs. Species-adjusted analyses retained the same three significant traits. BLINK sensitivity analysis of these traits detected genome-wide significance for NonACC_0 but did not reproduce the primary GEMMA lead locus. Under the prespecified +/-1.5-kb positional window, none of the 26 significant SNPs lay within or within 1.5 kb of an annotated gene; nearest-gene distances ranged from 2,465 to 340,885 bp. These results provide a conservative genomic framework for cold-response variation in Miscanthus and identify significant loci requiring independent functional and field validation.