Additional file 11: of Population genomics of rapid evolution in natural populations: polygenic selection in response to power station thermal effluents

Figure S6. Filtering impacts on coverage summary. (A) Density plot of reads per SNP per individual for the filtered (green) and raw (purple) datasets. (B) Density plot of non-zero reads per unique tag per individual for filtered dataset (purple) and the raw dataset (green). (C) Density plot of reads per SNP per individual for the filtered (green) dataset and the raw dataset excluding with SNPs where > 75% of individuals demonstrate zero reads (blue). Summary: There are large differences in coverage between our filtered (5.4 k) and raw SNP (314 k) datasets. The median read depth per SNP per individual of the raw dataset (panel A, purple) is 0 (filtered data median = 16, green). The mean read depth per SNP per individual of the raw dataset is ~ 4.4 (filtered data mean = ~ 26.3). While we base our filtering on missingness across SNPs and individuals (missing genotype calls results from less than 5 reads per unique 64 bp sequence tag within an individual), the majority of SNPs we filter out have zero reads in the majority of individuals, but coverage similar to our filtered dataset in the remaining individuals (panel B and C). (PDF 746 kb