Originally submitted as a Reader Response by Helen Wallace (helen.wallace@genewatch.org) on 12 April 2007:

Pharoah et al. [1] justify their search for common low penetrance susceptibility alleles for breast cancer on the basis that most of the familial clustering of breast cancer cases is as a result of inherited genetic variation rather than lifestyle or environmental factors. However, this statement is correct only if the three main assumptions of the classical twin study hold: no gene-gene interactions (epistasis); no gene-environment interactions; and the so-called ‘equal environments’ assumption for monozygotic and dizygotic twins. If any of these assumptions are incorrect, shared environmental factors may contribute significantly to the clustering of breast cancer cases in families.

The classical twin study solution (with an additive genetic model and no gene-environment interaction) provides only one of many possible interpretations of the twin and familial data for breast cancer: an infinite number of other models also could also fit the data [2]. The classical twin model result gives an upper limit to the genetic component of the variance of a trait in a given population, which reduces as the contribution of any gene-environment interaction increases, and also if epistatic terms begin to dominate the genetic model. Depending on the assumptions made about the importance of gene-gene and gene-environment interactions, shared environmental factors may partially or even largely explain the observed familial aggregation of breast cancer cases. Of the many possible alternative models, one solution consistent with the data could involve mutations in the BRCA1 and BRCA2 genes combined with a strong gene-environment interaction and a largely environmental explanation for the remaining unexplained clustering of breast cancer cases within families [2].

Pharoah et al. report that of the 710 common variants in candidate breast cancer susceptibility genes they studied, none achieved statistical significance. Their conclusion that larger sample sizes are needed is only appropriate if the common susceptibility genes they seek exist and contribute significantly to explaining the distribution of breast cancer cases in populations. If not, there is likely to be minimal benefit in conducting the expensive large-scale studies that they advocate [3].

References
1. Pharoah PDP, Tyrer J, Dunning AM, Easton DF, Ponder BAJ (2007) Association between common variation in 120 candidate genes and breast cancer risk. PLoS Genet 3(3): e42. doi:10.1371/journal.pgen.0030042.
2. Wallace HM (2006) A model of gene-gene and gene-environment interactions and its implications for targeting environmental interventions by genotype. Theoretical Biology and Medical Modelling 2006, 3:35 doi:10.1186/1742-4682-3-35.
3. Baker SG, Kaprio J (2006) Common susceptibility genes for cancer: search for the end of the rainbow. Br Med J 332: 1150-52.