Background

Epidemiological studies have shown that only about 20% of the familial clustering of breast cancer is explained by the known highly penetrant mutations in BRCA1 and BCRA2. We have set out to search for the genes for the remaining 80%. Twin studies indicate a predominant role of shared genes rather than a shared environment; the patterns of occurrence of breast cancer in families are consistent with a major polygenic component.

Methods

We have assembled a population based set of 5,000 breast cancer cases and 5,000 controls from the East Anglian population. We have simple clinical and epidemiological information, including family history, and samples of blood and paraffin embedded tumour.

We have used association studies based on single nucleotide polymorphisms, first with candidate genes and then in a genome-wide scan of 266,000 single nucleotide polymorphisms, to search for the putative predisposing genes. We have as yet searched only for common variants (frequency >5%).

Results

We have modelled the effects of polygenic predisposition in the East Anglian population, and have shown that the model predicts a wide distribution of individual risk in the population, such that half of all breast cancers may occur in the 12% of women at greatest risk.

Both the candidate gene-based and genome-wide scans have provided provisional identification of a number of novel susceptibility genes, and these are currently being confirmed by a world-wide consortium of independent laboratories totalling 20,000-plus cases and controls. No single gene so far identified contributes more than 2% of the total inherited component, consistent with a model in which susceptibility is the result of a large number of individually small genetic effects.