Skip to navigation Skip to main content Skip to footer

Approved research

Phasing human populations

Principal Investigator: Professor John Hickey
Approved Research ID: 21413
Approval date: July 21st 2016

Lay summary

Genotype phasing and imputation is an important step in many genetic studies. Genotyping technology normally produce unphased genotype and do not provide the underlying haplotypes. Inferring these haplotypes can improve many genetic studies as they provide additional information. Accurately inferred haplotypes can also improve genotype imputation. Genetic studies benefit from genotype imputation as the inclusion of imputed markers increases the power of many of the tests involved in such studies. We propose to investigate whether out genotype phasing and imputation software can outperform existing methods when imputing on human datasets. Our intention is to improve the available phasing and imputation methods for human datasets and so maximize the benefit to researchers investigating the association between health and genetics. We have developed AlphaImpute, a tool for phasing and imputing missing genotypes. To date AlphaImpute has been successfully used on animal data but we believe our methods have the potential to outperform existing methods on human data. By optimising our software for use in a human context we hope to improve it's performance further and, by doing so, we will maximise it's contribution to human health genetics research. Our software, AlphaImpute, can phase and impute large numbers of genomes with high accuracy. We have tested AlphaImpute on many animal species and now wish to test it on a human dataset. We propose that we would first test the performance of AlphaImpute on the genetic data of the full cohort. This testing is likely to identify areas where we could improve our method to better account for the unique challenges of a human dataset. We propose that we would test the performance of AlphaImpute on the genetic data of the full cohort. By using the full cohort we would maximise the usefulness of our imputation to other researchers.