Prioritizing sequence variants from cancer families

This project will consider a large set of genetic variants shared among family members affected with a rare cancer. To narrow the focus of future investigations, we will explore ways to prioritize variants that are predicted to be functionally relevant and shared with distantly related affected relatives but not with unaffected relatives, while taking into account the rareness of the particular cancer subtypes in the family and the age and gender of the cases. For example, a family with two affected members whose relationship dictates that they would be expected to share half of their variants by chance is of lower practical value than a family with three affected members whose relationships dictate that they would only be expected to share one out of every 16 of their variants by chance. As the variants identified in the second family have a lower prior probability   of being due to chance, they should be weighted more heavily than the variants identified in the first family when prioritizing the combined list of variants for further investigation. In this project, we will explore statistical approaches to integrate all these different sources of information -- functional predictions, relationship information and epidemiologic information on population disease incidence -- when prioritizing sequencing variants for further study.