3.2 | Genome-wide association analysis
Our genome-wide association study found 25,854 candidate SNPs with strong associations with virus exposure (BF > 50), with the PCA based on candidate loci revealing clear patterns of genetic structuring between locations varying in historical AVG exposure (Figure 2). Estimates of LD were high at all locations (meanr 2 = 0.61 ± 0.01 SD ) indicating non-random association of alleles, while comparisons ofr 2 between virus affected and unaffected stocks did not significantly differ (P > 0.05). Analyses suggest a genome wide patterns of selection, with regression analyses indicating strong linear relationship between number of candidate loci and scaffold length (R 2 = 0.83; Figure 3a). However, scaffolds QXJH01000030.1 and QXJH01000212.1 exhibited a higher number of candidate loci relative to scaffold length (deviating from the linear distribution), with candidate SNPs comprising 0.0062% and 0.0108% of the total scaffold nucleotides, respectively. Pairwise linkage among SNPs across the entirety of these scaffolds was high (D ’ ≈ 1), with the detection of large haplotype blocks indicating large sections of the genome linked to virus exposure (Figure 3b).