Statistical techniques to infer ancestral history and selection using genome-wide DNA

Thursday 18 October 2018, 1.00PM

Speaker(s): Dr Garrett Hellenthal, UCL

I describe new statistical approaches to infer the ancestral history of populations and individuals, using haplotype information from dense SNP data. I demonstrate how these techniques can describe and date admixture events involving historical interactions among groups spanning thousands of years, as well as shed light on the time spans over which groups became isolated from one another. I also describe new methods to test for genetic loci affected by selection using both allele frequency and haplotype data. These novel techniques can identify the loci and populations involved in selection, enable the joint testing of SNP sets (e.g. in pathways), and account for both demography and variation in data quality across loci and samples. We showcase these models with applications to several world-wide populations, incorporating DNA from both ancient and modern humans.

Location: K/018