Proceedings of the 2013 annual meeting of the Netherlands Epidemiology Society
Volume 1 Issue S1 Abstract 53
F.H. Van der Baan, Netherlands Cancer Institute, Amsterdam, The Netherlands
C. Canisius, Netherlands Cancer Institute, Amsterdam, The Netherlands
W. Wessels, Netherlands Cancer Institute, Amsterdam, The Netherlands
M. Rookus, Netherlands Cancer Institute, Amsterdam, The Netherlands; CIMBA
In pathway analyses the unit of analysis is not an individual SNP or gene, but a pre-specified combination of genes, based on prior biological knowledge. These analyses can complement single-SNP analyses as most phenotypes are expected to result from multiple genes. We selected 2242 SNPs from a meta-analysis of 32 GWAS for age of menarche. As in the general population early menarche is an established risk factor for breast cancer, we aim to conduct pathway analyses on these menarche SNPs with risk of breast cancer in BRCA1/2 mutation carriers.
Based on the SNPs, we selected 100 human pathways from four different databases: KEGG, PID, BioCarta and Reactome. We chose the adaptive rank truncated product (ARTP) method as analysis approach, since this is a self-contained test, suitable for both SNP-based and gene-based pathway analyses, and it uses a permutation procedure to estimate the pathway p-value. Data from 14,797 BRCA1 and 7,993 BRCA2 mutation carriers from the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA) were used.
The ARTP method resulted in multiple gene sets enriched with associations with breast cancer in BRCA1/2 mutation carriers. Currently we investigate these biological pathways in more detail.
Performing pathway analyses in addition to single-SNP analyses is important to reveal combinations of SNPs or genes that are associated with a phenotype, even more as they may have too subtle effects to be detected individually. The incorporation of biological knowledge may subsequently be helpful in the interpretation of the results.
Published: 06 Jun, 2013