Date/Time
Date(s) - 26/03/2013
1:00 pm - 2:00 pm
Location
SGDP Seminar Room
Category(ies)
Dr. Chris Wallace from Cambridge University
https://www-gene.cimr.cam.ac.uk/staff/wallace/
Abstract:
Integration of data from genomewide single nucleotide polymorphism
(SNP) association studies of different traits has the potential to
reveal the shared and disparate genetic mechanisms underlying related
but distinct traits. For example, comparing disease and eQTL
association scans could reveal the underlying causal genes and
relevant tissues in cases where the effect of disease associated
variants are mediated by variation in gene expression. I will discuss
the underlying statistics, including reasons behind the poor type 1
error rate control of some published methods and show how principal
components and Bayesian model averaging can be applied to overcome
this.
Data from the autoimmune thyroid diseases Graves’ disease and
Hashimoto’s thyroiditis are used to illustrate the methods discussed.
Overall, autoimmune thyroid diseases share genetic aetiology most
closely with rheumatoid arthritis and type 1 diabetes. In individual
regions associated with both diseases, the causal variants are shared
and predispose equally to either disease. In contrast, in five out of
six regions which have been significantly associated with one disease
and not the other, the lack of evidence in one disease represents
genuine absence of association rather than lack of power.