Finding genes via shared ancestry of genome

Elizabeth Thompson, University of Washington, visiting St. Andrews. Part of the statistics seminar series.

The  relatedness of individuals  is reflected in the close similarity of segments of their genomes that are descended from shared common ancestors.  The descent of genome segments also leads to population-level dependence in DNA observed at contiguous genome locations.  Modeling this dependence in the DNA both among individuals and across genome locations is key to using modern genomic data in the mapping the locations of DNA that contribute to a quantitative trait. Traditionally, genetic marker data were sparse, and defined pedigree relationships were used to infer the shared descent (IBD) leading to correlations among relatives.  With modern genetic data, this pedigree prior is unnecessary, and both local and genome-wide IBD may be estimated from genetic data alone.   However, even pairwise local IBD is not perfectly estimated, and levels of IBD vary due to chance events of meiosis.  In population-based mapping analyses, adjustment for the background level of IBD is essential. We present current methods for model-based IBD estimation across the genome, and also show how Kullback-Leibler information can be used to provide a more reliable assessment of the meaning of a linkage signal.

Elizabeth Thompson, University of Washington, visiting St. Andrews