Challenges in the identification and use of rare disease-associated predisposition variants

The case for identifying rare, disease-associated germline variants rather than undertaking larger genome-wide association (GWA) studies for common variants is increasingly being advocated. I agree with the importance of identifying rare variation in human disease, but believe more thought needs to be put into the limitations of such enterprises before advocating a dramatic departure from the GWA approach. In this paper, I discuss some of the main challenges in identifying rare disease variants with modest effects in disease, including: the overoptimistic expectations about their effects; the need for very large studies needed to prove, beyond doubt, statistical associations; the problems associated with private variants, including the need for functional studies; the difficulties in prioritisation of candidates for further validation and the issues related to the accurate estimation of the risk associated with individual rare variants. The rare variant approach is very promising, but it remains largely untested in comparison with the proven and successful GWA approach. Both strategies must continue to be pursued in parallel and their advantages and pitfalls must be considered without excessive scepticism or expectation.