Poster Sessions

BIOINFO-24

Idan Menashe

I. Menashe, D. Maeder, S. Bhattacharjee, M. Rotunno, P.S. Rosenberg, N. Chatterjee

Pathway Analysis of Breast Cancer Genome-Wide Association Study Highlights Three Pathways and Two Canonical Signaling Cascades

Genome-wide association studies (GWAS) focus on relatively few highly significant loci while less attention is given to other genotyped markers. Employing pathway analysis to existing GWAS data may shed light on relevant biological processes, and illuminate new candidate genes. We studied the NCI’s CGEMS breast cancer GWAS (1145 cases and 1142 controls) and retrieved pathways from three resources: KEGG database, BioCarta, and the NCI Protein Interaction Database (PID). Genes were represented by their most strongly associated SNP, and pathways enriched with gene-based association signals were determined by a weighted Kolmogorov-Smirnov (KS) procedure. Of the 421 pathways included in our study, three pathways (‘Syndecan-1-mediated signaling ‘, ‘Signaling of Hepatocyte Growth Factor Receptor’ and ‘Growth Hormone Signaling’) were highly enriched with association signals (PKS < 0.001, False-Discovery Rate = 0.118). Using clustering analysis, we found that pathways containing the genes GRB2, SOS1, HRAS, RAF1, MAP2K1 and MAPK3 and pathways containing the genes WNT1, FZD1,DVL1, GSK3B, APC and CTNNB1, were more likely to be enriched with association signals than expected by chance (P = 0.0051 and P = 0.0365 respectively). Our results suggest that genetic alterations associated with these three pathways and two canonical signaling cascades may underlie breast cancer susceptibility.