Signature AIMs are ancestry-informative SNPs that exhibit complete, or near-complete specificity for a single population group, characterised as a population-specific allele occurring at an informative frequency in the target population (≥0.1) but at extremely low or zero frequencies in all others (≤0.001). Signature AIMs were identified in 1000 Genomes, HGDP-CEPH and Almarri’s Middle East [1] whole-genome sequence SNP datasets and validated with reference to geographically broader based but smaller sample size Simons Foundation and Estonian Diversity Project population data.
From a compilation of several thousand SNPs, varied numbers of loci were combined to give an average 20-22 population-specific alleles per individual with an origin in Africa, Europe, East Asia, South Asia, America, Oceania, Middle East or North Africa. An AmpliSeq MPS multiplex of 328 Signature AIMs was constructed from an initial input of 342 candidates (96% assay conversion rate), applying a ‘serial maximum amplicon length threshold’ system to create a panel design of 280 SNPs <140 bp; 17 <175 bp; 18 <275 bp and 13 <375 bp.
We initially expected simple Signature AIM specific-allele counts to provide a secure method to assign population-of-origin, even in parental admixture samples where ~10 specific alleles could be expected from both contributing co-ancestries. However, allele counting alone is not sufficiently precise for real-world samples - notably those from continental margins, where average specific-allele counts can drop to as low as 40% of the expected average number. We are developing a technique where the observed number of non-specific alleles in each population is used to adjust for spurious counts in individuals where gene-flow at continental margins has disrupted the degree of specificity shown by the selected SNPs.
[1] M.A. Almarri, et al., The genomic history of the Middle East, Cell 184 (2021) 4612-4625.