Geneticists and neuroscientists often differ dramatically in their interpretation of the statistical significance of genetic association studies, and this is particularly true when fMRI phenotypes are being assessed in relation to one or a few candidate polymorphisms. This presentation will review the assumptions that underlie these differences, with particular emphasis on the concept of genome wide significance, and how this concept has been applied in genetic mapping of human diseases and other traits. Thresholds for genome wide significance were first applied to genetic linkage studies beginning in the 1980s and are now being applied in well powered genome wide association studies (GWAS). GWAS of numerous traits are now being conducted using up to one million single nucleotide polymorphisms (SNPs) in samples of up to tens of thousands of individuals, and have already led to discovery of more than 100 novel genetic loci, for which genome wide significance has been achieved in at least two independent studies. In the era of GWAS it is challenging to design reasonably powered genetic association studies for measures, such as those obtained in fMRI, where it is typically not feasible to assess large numbers of individuals. In this presentation I will discuss one approach to this problem, which we plan to implement in the Consortium for Neuropsychiatric Phenomics at UCLA.
Audio (MP3 File, Podcast Ready)
Copyright. All Rights Reserved.