Meeting Abstract

S4-2.3  Saturday, Jan. 5  A “forward genomics” approach links genomic and phenotypic evolution in a clade of related species HILLER , M; BEJERANO , G*; Stanford University; Stanford University bejerano@stanford.edu

Genotype to phenotype association is a holy grail of the genomic era, hampered by the lack of clear mappings between the millions of genomic changes and thousands of trait differences apparent even when comparing closely related species such as human and chimpanzee. Efforts to link DNA base pair changes to whole organism phenotypes have recently focused on experimentally mapping genomic regions involved in a given trait or testing genomic regions that show accelerated changes between lineages. Here we introduce a computational “forward genomics” strategy to detect phenotype – genotype associations by matching a phylogenetic pattern of trait evolution with a corresponding pattern of orthologous genomic regions evolution. Simultaneously searching dozens of mammalian genomes we are able to correctly associate individual genes with the phenotypic traits to which they contribute. We show that our method is robust to missing phenotypic data, and applicable for both discrete and continuous, monogenic and polygenic traits. Using simulation studies, analysis of existing phenotype surveys and the coming availability of genomes of many additional species we show that “forward genomics” can be applied to many phenotypes, including those relevant for human evolution and disease. A portal allowing researchers to query their phenotypes of interests for matching genomic regions is developed at http://bejerano.stanford.edu/phenotree/