**Background:** Wheat (Triticum sp.) underwent severe genetic bottlenecks from polyploidization, domestication, and the transition to free-threshing grain, yet it adapted to a vast range of environments. Understanding the genetic basis of this adaptation and the origins of introgressed segments is important for breeding. A previous study (He et al.) identified a large introgression on chromosome 4A from wild emmer wheat, but the authors hypothesized it might instead derive from the Norin-10 semidwarf cultivar lineage.
**Methods:** Two SNP datasets were used: (1) exome capture data (~8.7M SNPs, filtered to ~113K SNPs, 921 accessions) from TriticeaeToolbox, and (2) whole-genome resequencing data (~100M SNPs) from Genome Variation Map. Population structure was assessed via ADMIXTURE, PCA, and neighbor-joining phylogeny. Genome scans for selection signatures used PCADAPT. Genome-environment association (GEA) was performed with LFMM2 using latitude, temperature, and earth skin temperature data for 113 landraces with coordinates. Candidate genes within 200 kb of significant loci were annotated via WheatGmap, Phytozome, and literature. To trace the 4A introgression origin, FST, PhiPT, and G-tests were computed for the most significant 5000 SNPs in the 4A region across diverse wheat types and wild relatives.
**Key Results:** Population structure revealed eight subpopulations (K=8) in the full dataset and seven in the landrace subset, largely reflecting geography. The largest genetic differentiation (FST = 0.34) was between Former Soviet Union (FSV) and Old World (OLDWP) subpopulations. Genome scans identified a highly differentiated region on chromosome 4A spanning ~174–465 Mb, plus outlier loci on chromosomes 2A, 5A, 6A, 7B, 1B, 4B, 5D, 7A, and 7D. Within the 4A region, 62 extreme outlier SNPs (p < 1 × 10⁻⁷⁵⁰) dichotomized the population into two groups: G1 (dominated by haplotype 1, 560 individuals) and G2 (dominated by haplotype 2, 115 individuals). G2 included ~130–140 of 347 landraces and 250–260 of 921 total accessions, with higher frequency in Mediterranean and Middle East regions. Haplotype analysis identified eight haplotypes; Hap1 and Hap2 together represented 675 of 921 accessions. GEA identified nine loci associated with latitude or temperature (FDR < 0.05), including Chr7A:118327401 (p = 1.63 × 10⁻⁸) and Chr2A:344449220 (p = 4.44 × 10⁻⁷) associated with latitude and temperature, respectively. The 4A region harbored >25 seed dormancy genes, >30 flowering-time genes, and >50 abiotic stress response genes. Candidate adaptive genes included SIGA, RBR1, CIPK, AGL30, SIZ2, and GCR1. Crucially, FST and PhiPT analyses showed hexaploid cultivars and durum wheat had zero differentiation from wild emmer at the 4A region (FST = PhiPT = 0), but the G-test (10,000 Markov chains, 5000 iterations) strongly differentiated cultivars from wild emmer while cultivars remained undifferentiated from other cultivated types (Indian dwarf, Tibetan semi-wild, Xinjiang wheat, etc.). Ancestral proportion analysis (K=5) showed >95% shared ancestry between cultivars and wild emmer, but also with these other cultivated wheats. The landrace-spelt group formed a distinct clade with long branches, concentrated in the Mediterranean region.
**Clinical Implications:** While not a clinical study, this research has significant implications for crop science and food security. Identifying adaptive genomic regions and their origins can guide wheat breeding for climate resilience, particularly for traits like drought, heat, and cold tolerance, flowering time, and seed dormancy. The finding that the 4A introgression likely came from cultivated wheat rather than wild emmer refines our understanding of wheat breeding history and highlights the value of landraces and heritage cultivars as genetic resources. This can inform conservation strategies and the development of better-adapted wheat varieties for diverse environments.