Analyses of genomic variation
Genomic differentiation between grossly normal and wasting sea stars was very low based on Fst estimates. Across the final SNP dataset (71,784), Weir & Cockerham’s Fst had a median value of 0.00314, and nearly 45% of loci had Fst = 0 (Table S2). Moreover, while 362 loci had moderate Fst values (≥ 0.1), no locus showed significant differentiation after FDR adjustments; the lowest adjusted P-value was 0.172 (Figure 1). Outlier tests with Fst using Bayescan also showed no evidence of selection in any locus in our dataset (Figure S3).
DAPC analyses showed modest separation between apparently normal and wasting groups (Figure 2A). Based on loading values from the DAPC, we identified 18 SNPs across 10 chromosomes contributing most to the differentiation between the two groups (Figure 2B, Table S3). Allele frequency differences across these loci ranged from 0.014 - 0.253 (Table S3). Using a 30-kb window centered at each of these SNP positions, we detected 34 protein-coding genes predicted by the genome annotation from Ruiz-Ramos et al. (2020). BLAST of these protein sequences against the Uniprot/Swissprot databased revealed that 30 of them have predicted products with known functional annotation (Table S4). Chromosomes 3 and 8 harbored the most genes linked to these SNPs (seven genes each); in chromosome 8, one SNP was linked to three genes, and the others to two each (Figure 3).
We assessed whether the 18 SNP loci identified as outliers from our DAPC analyses overlapped with haplotypes from Schiebelhut et al. (2018) that were the most discriminatory between post- and pre-SSWS adults in their samples from California. For this, we compared the 30-kb ranges encompassing our SNP outliers to the haplotype coordinates from Schiebelhut et al. (2018). We detected no overlap in the two sets of outliers, with the shortest distance detected as ~257 kb (Table S5).