My name is Professor Jonathan Marchini.
I'm a professor of Statistical Genomics
at the Department of Statistics, University of Oxford.
I will be giving a lecture on imputation in "Genome-Wide Association Analysis".
The aim of this talk is to introduce the idea of
genotype imputation for genome-wide association studies.
I will start with a short overview of what genotype imputation is and then
we'll give a quick summary of the basic idea behind how imputation works.
I will then describe one of the first methods of
genotype imputation post called IMPUTE v1.
Many of the state-of-the-art algorithms currently
available for genotype imputation build upon this approach.
I'll then describe how imputed data can be used to test for association and illustrate
these methods on real genome-wide association data
from the Wellcome Trust Case Control Consortium.
I will discuss the issue of quality control measures for imputed genotypes
and the importance of alignment of strand between reference panel and GWAS samples.
I'll illustrate how imputation can facilitate
meta-analysis of genome-wide association studies,
and then discuss the factors that influence accuracy.
I'll talk about the 1000 Genomes Project,
which create a world-wide reference panel for
imputation and illustrate the accuracy of imputation using this resource.
It's now routine for GWAS samples to be phased before imputation.
This means that the underlying haplotypes of each sample are
estimated using statistical methods before imputation.
I will describe this process and also some of
the most accurate methods for phasing which are available.
And then finally, I'll discuss the Haplotype Reference Consortium,
which has recently constructed the haplotype reference panel approximately
65,000 haplotypes for using genotype imputation.
I'll describe the accuracy of using this resource.
How researchers can access it.
A typical genome-wide association study consists of