Executive Summary

Cancer Genetic Markers of Susceptibility (CGEMS) Project

The Cancer Genetic Markers of Susceptibility (CGEMS) project is an NCI enterprise initiative to conduct whole genome association studies to identify genes giving rise to risk of prostate and breast cancer. The project is designed to capitalize on new knowledge of human genetic variation and technical advances in ultra-high-throughput genotyping. Single nucleotide polymorphisms (SNPs) are the most common form of human genome variation. Information from the International HapMap Project, and from extensive resequencing of representative genomic regions, suggests that most of the roughly 10 million SNPs with a minor allele frequency greater than 5% (common SNPs) occur in genomic segments in which these SNPs are highly correlated with each other; in other words, they are in linkage disequilibrium, LD. Thus, a carefully selected subset of SNPs will represent a much larger number of common SNPs, permitting cost-effective whole genome scans for detection of SNPs that have different prevalence in cancer cases than controls. Estimates based on the International HapMap Project Phase 2 data indicate that at minimum 550,000 carefully chosen SNPs would be required to conduct a dense whole genome SNP scan. The number of SNPs required could reach 1 million for populations in which there is more disintegration of LD.

CGEMS is designed to conduct whole genome scans in nested case-control studies of prostate cancer and breast cancer from large ongoing population-based cohort studies. The detection of genes or regions in the genome that are associated with risk for prostate and breast cancer will be useful for understanding causal pathways, as well as providing tools for risk stratification in the development of preventive interventions. Prostate and breast cancers were selected because they met the following criteria:

They are major causes of cancer death and suffering,
There is strong prior evidence of a substantial inherited genetic component of risk,
Sufficient high-quality epidemiologic studies exist to assemble the requisite samples and data.

Each whole genome scan will analyze approximately 1,200 cases and 1,200 controls. Following the scan, sequential replication studies will be performed to validate true positive associations. The very large number of SNP comparisons in a whole genome scan is likely to give rise to many false positive signals. Therefore, the CGEMS study design is to assess the top candidate SNPs from the whole genome scans (15,000-20,000) in follow-up case-control studies. The follow-up samples will be nested within cohorts or within high-quality population-based case-control studies (see figure). SNPs that are significantly associated with cancer risk in sequential replication studies will be made rapidly publicly available through the NCI caBIG database. We intend to conduct these phased studies in an accelerated program over three years. The findings of CGEMS will require biological characterization by the cancer research community in order to develop possible clinical applications.

The CGEMS study design has the following advantages:

It will conduct scan and replication in existing cohort and case-control samples and thus, leverage the resources committed to population-based studies in cancer etiology.
It will contribute practical information for the conduct and analysis of whole genome scans, including strategies for replication of results with significant p values. This should inform future studies and accelerate the discovery of susceptibility genes for a wide range of cancers.
The initiative will inform the strategies for reducing costs through economies of scale across technologies and vendors, and through sharing of information. Integral to this process is the creation of a caBIG-compatible infrastructure that ensures immediate public access to data.

CGEMS is a strategic partnership between intramural and extramural groups that are joining forces to conduct large-scale studies in high quality epidemiologic studies. This approach will ensure use of the highest quality sample sets and will facilitate diffusion of results across the intramural and extramural research communities. The project is being coordinated through the NCI’s Division of Cancer Epidemiology and Genetics, the Core Genotyping Facility, and the Office of Cancer Genomics.

Studies that are participating in the CGEMS prostate cancer program include:
American Cancer Society CPS-II
ATBC Study
European Prospective Investigation on Cancer (EPIC)
French Prostate Cancer Study (CeRePP)
Health Professionals Follow-up Study
MultiEthnic Cohort Study
Physicians' Health Study
PLCO Study

Studies that are participating in the CGEMS breast cancer program include:
American Cancer Society CPS-II
European Prospective Investigation on Cancer (EPIC)
MultiEthnic Cohort Study
Nurses' Health Study
Polish Case-Control Study
PLCO Study
Women's Health Initiative

CGEMS Co-Leaders
Stephen J Chanock, M.D.
Robert N. Hoover, M.D., Sc.D.
David J. Hunter, M.B.S.S., Sc.D.
Gilles Thomas, M.D., Ph.D.

Executive Directors
Joseph F. Fraumeni, M.D.
Daniela S. Gerhard, Ph.D.