Population structure analysis and SNP-based heritability estimation
A quality control analysis was conducted on the materials using Plink (Purcell et al., 2007). Initially, SNPs with a minor allele frequency (MAF) < 0.05 were removed. Subsequently, materials with SNP missing rates > 0.1 were excluded. Finally, after filtering the original SNPs (Tang et al., 2021), a final set of 8,503,071 SNPs was obtained, with a median MAF of 0.21. In addition, Plink was used to calculate the genomic relationship matrix (GRM) among the 505 accessions. Principal component analysis (PCA) was then performed based on this matrix. In the resulting PCA plot, the accessions were differentiated into subgroups reflecting their ecological types. Our materials included five ecological types: semi-winter 1 (SW1), semi-winter 2 (SW2), semi-winter mixed (SWM), spring (SPR) and mixed.
Using the GRM above, the SNP-based heritability was calculated with SOC as the phenotype. This analysis was conducted for each environment using GCTA (Yang et al., 2011), with the recommended parameters ‘–bfile –make-grm –make-grm-alg 0 –out; –grm –pheno –reml –out’.