Mini-haplotypes as lineage informative SNPs and ancestry inference SNPs

Abstract

We propose that haplotyped loci with high heterozygosity can be useful in human identification, especially within families, if recombination is very low among the sites. Three or more SNPs extending over small molecular intervals (<10 KB) can be identified in the human genome to define miniature haplotypes with moderate levels of linkage disequilibrium. Properly selected, these mini-haplotypes (or minihaps) consist of multiple haplotype lineages (alleles) that have evolved from the ancestral human haplotype but show no evidence of recurring recombination, allowing each distinct haplotype to be equated with an allele, all copies of which are essentially identical by descent. Historic recombinants, representing rare events that have drifted to common frequencies over many generations, can be identified in some cases, they do not equate to frequently recurring recombination. We have identified examples in our data collected on various projects and present eight such mini-haplotypes comprised of informative SNPs. We also discuss the ideal characteristics and advantages of minihaps for human familial identification and ancestry inference, and compare them to other types of forensic markers in use and/or that have been proposed. We expect that it is possible to carry out a systematic search and identify a useful panel of mini-haplotypes, with even better properties than the examples presented here.

Figure 1

Haplotype frequencies (proportional to colored bar lengths) in 45 populations from around the world. Each mini-haplotype consists of three LISNPs except for KRAS which has four LISNPs. Populations are ordered geographically on the x axis as in Table 1.

Figure 2

A scatterplot of each of the eight minihaps in each of the 45 populations plotting the heterozygosity and the percentage of individuals with an unambiguous genotype. The eight different loci are indicated with different symbols, as indicated. One PAH point (73% heterozygosity and 10% resolvable) falls outside the range visualized.

Figure 3

Frequency bar graphs for 3 of the 24 ‘haploblock' haplotypes identified by Ge et al using HapMap information. Compare to bar graphs in Figure 1 for the eight minihaps.