Mammographic density (MD) is one of the strongest established risk factors for breast cancer and has an estimated heritability of over 60%. Genome-wide association studies (GWAS) to date have explained only a small fraction of the heritability. We propose to combine gene expression and network information with GWAS data to augment the power to discover MD genes, and to gain insights into the biological mechanisms underlying MD and its association with breast cancer risk.
Our specific aims are:
Aim 1 Conduct a transcriptome-wide association study (TWAS) of MD in 24,192 women, and replicate findings in the Marker of Density (MODE) consortium.
Aim 2 Conduct network analyses to discover gene sets (network modules) acting jointly on MD and elucidate the underlying biological pathways. We will construct tissue-specific gene co- expression networks using transcriptome data in normal human breast tissue samples from GTEx, and develop new statistical methods to integrate these tissue-specific networks to boost the power and accuracy of gene expression-based association tests.
Aim 3 Evaluate associations of MD genes and modules with breast cancer risk using summary statistics from international breast cancer consortia, and individual-level GWAS data from 60K women in Kaiser's Research Program on Genes, Environment and Health (RPGEH) and independent replication set of 45K women in public GWAS data repositories. The proposed approach is expected to have substantially higher power than single-variant GWAS approaches because it rationally combines information, first across multiple SNPs using gene expression levels as an intermediary, and second across multiple genes using gene co-expression networks to model the correlation and interaction among genes. Moreover, gene- and network-based associations naturally provide a biological context, and are more easily interpreted than single SNP-based associations. The proposed research is innovative because we will develop new methods and a rational framework, based on gene expression and co-expression, to conduct gene-based and network-based association tests, which may be applied to study other traits. The results will improve our understanding of the genes and biological mechanisms underlying MD and its association with breast cancer risk, and may lead to the development of more effective therapies to prevent breast cancer.
High mammographic breast density is the most common and one of the strongest risk factors for breast cancer, but the underlying biology remains poorly understood. This study will combine functional data from gene expression and co-expression networks with GWAS data using new approaches to discover density genes with greater power. The results will shed light on the biological mechanisms underlying the association of breast density with cancer risk, and may lead to improved preventive therapies.
