I, too, was a little surprised by the glibness of Debbie’s description of her group’s approach to Booleanize data that would seem to be much richer if not so transformed. However, I have a tremendous amount of faith in Debbie and her collaborators and her track record. She was giving a very short talk (on the order of about 20 minutes) and most likely needed to quickly describe one easy-to-understand approach. I feel very confident that they will analyze their data in a robust manner, taking advantage of multiple statistical genetic approaches to analysis. There are some issues of multiple test correction and maximizing the power of their study that might benefit from an up-front declaration of a Boolean approach as one primary outcome variable.

With respect to whether whole genome studies should consider the results of previous linkage studies, I would recommend that most not include data from previous studies in a primary analysis, but that a secondary analysis be performed with those results. Long conversation, but in short: “looking under the lamp post.”

Stacey Gabriel is looking for a variants with assumed large odds ratio, with certain functional characteristics, within certain genes. People have previously done association and linkage studies that do the same thing, and lots of times nothing has been found; she should look to see whether her study falls into the realms of one of those.