Loading...
Imputation of ancient human genomes from under-represented Native American populations
Hay, Savannah Beaty, Kristine G. Tackney, Justin Macleod, Ruairidh Barrett, Christopher Thomas, Mark G. Garcia, Obed Jorgensen, Kelsey O'Rourke, Dennis Raff, Jennifer
Hay, Savannah
Beaty, Kristine G.
Tackney, Justin
Macleod, Ruairidh
Barrett, Christopher
Thomas, Mark G.
Garcia, Obed
Jorgensen, Kelsey
O'Rourke, Dennis
Raff, Jennifer
Citations
Altmetric:
Abstract
Historically, imputation to infer missing genotypes in genetic datasets was constricted to use on contemporary genetic data. This limitation was based on two concerns: the inferring of past genetic diversity using contemporary reference panels and the very low sequencing coverage of most ancient genomes. However, studies in recent years have shown that established imputation methods are suitable for ancient samples with depth of coverage as low as 0.5-0.75x, when using the 1000 Genomes Project cohort as a reference panel. This panel has been used to successfully impute genomes from less represented populations, including Native Americans. There is, however, a limited number of Native American individuals, from particular geographic regions, in this panel (n=347). With this limited number, it is possible that some imputation targets are significantly diverged from the haplotypes present in the panel. To explore the suitability of genotype imputation upon targets with genetic ancestry sources not well-represented by modern reference variation data, we considered several approaches for genotype imputation and phasing (including GLIMPSE) with the 1000 Genomes Project Phase 3 reference panel. In particular, we describe the outcomes of this approach using an ancient genome, sequenced initially to 11X autosomal depth but then resampled at lower depths, from a 770–380cal BP Unangax individual, resampled at lower depths of coverage. The populations of the Aleutian Islands have a distinct population history from the Central and South American groups that make up the Native American component of the 1000 Genomes Project. To mitigate genotype missingness, we determine the value of using additional reference panels, such as the imputation server provided by TOPMed. We also test the power of using a smaller reference panel created from less genetically diverged populations. We present additional steps researchers can take to address this concern and improve imputation accuracy when working with genetically under-represented groups.
Description
This is the poster from a presentation given at the American Society of Human Genetics held in Boston, MA on 10/16/25.
Date
2025-10-16
Journal Title
Journal ISSN
Volume Title
Publisher
University of Kansas
Files
Loading...
HayS_2025.pdf
Adobe PDF, 1.4 MB
Research Projects
Organizational Units
Journal Issue
Keywords
Population genetics, Imputation, Bioinformatics