Identifying and accounting for ascertainment bias in the Ae. aegypti SNP-chip using whole genome sequencing

Aedes aegypti is the primary vector of relevant arboviruses to humans. Genetic population studies are essential to inform vector surveillance and control programs worldwide. The aedes axiom1 SNP-Chip is capable of quickly genotyping ~46k SNPs across the Ae. aegypti genome at a moderate cost and with minimal bench-work required. However, SNP-Chips face frequent criticism regarding ascertainment bias and its effect on demographic inferences. We explored the ascertainment bias of the Ae. aegypti SNP-Chip in order to assess the accuracy of its use for population genetic studies. We compared SNPs derived from genotyping the same individual mosquitoes, representing African and outside-Africa populations, with both the SNPs-Chip and by a low-depth Whole Genome Sequencing (WGS) to determine concordance. There was an overall genotype concordance of 68.7% and heterozygote concordance of 66.2% across samples, with the SNP chip showing a remarkable variant allele sensitivity, recovering 94.7% of variants detected by WGS. Analysis of random subsets, containing 46K to 208K SNPs (DP>10), indicate that the chip yields a similar nucleotide diversity and overall observed heterozygosity than WGS, but similar relative divergence between individuals and consistent clustering. Differences in relative genetic diversity spectrum are likely driven by the exclusive representation of variable sites in the SNP-Chip and the active inclusion of diagnostic variable sites across world populations. Our results suggest that analyses involving nucleotide patterns and the allelic variation spectrum must be carefully assessed prior to using the entire SNP-Chip dataset, but confirms SNP-Chip suitability for genome population studies involving local demography and structure.