YKBS

**2014년 2월 YKBS 정기세미나**

 

연사: 최진명 박사님 Department of Neurology

일시: 201년 2월 20일 (목) 오후 6시

장소:HOPE216

 

Title: Uncovering the genetic architecture of complex traits using network approaches

 

Over the last decade, genome wide association studies (GWAS) have uncovered thousands of genetic variants influencing complex traits affecting all organ systems. However, it has proven difficult to infer which genes these variants are perturbing, so the challenge of uncovering underlying biological mechanisms remains elusive. We do not yet understand the genetic architecture of traits: do variants perturb entire pathways or other groups of interacting genes? How big are these groups? Are many distinct pathways involved or is a single molecular system sufficient to alter traits? We have developed a robust and flexible framework to address these questions by detecting groups of interacting genes perturbed by associated variants. We have previously observed that genes encoded in disease risk loci tend to interact, suggesting that risk variants modulate susceptibility by perturbing gene networks representing pathways [Rossin et al PLoS Genetics 2011]. Here, we project GWAS scores onto protein-protein interaction networks and use robust network clustering algorithms to look for regions of the overall interaction network enriched for disease association signal, indicating the presence of an interacting set of genes modulating susceptibility. Within this framework, we investigate several genetic architecture models by simulation: scenarios where either a single or multiple groups of interacting genes modulate risk versus no aggregation of signal in gene groups. We are able to recover both true positive and false negative associations, increasing the heritability explained by GWAS. Surprisingly, we also find that in models of multiple distinct susceptibility pathways, these are detected as a single large connected gene network. This appears to be a property of gene interaction networks and suggests that recent reports of large sets of interacting genes underlying disease susceptibility are in fact capturing multiple biological pathways. In two disease GWAS meta-analyses (Crohn’s disease and multiple sclerosis) we find gene networks of size ~100 enriched for genetic risk, consistent with an architecture of cumulative genetic burden on molecular pathways. We also find that subsets of these components are expressed in different tissues, indicating that both diseases involve multiple pathways, not a single large group of genes. We are currently extending this framework to simultaneously consider both diseases and investigate pleiotropic effects in these networks.

 

회원님의 많은 참여 부탁드립니다.