Reader Comments

Post a new comment on this article

Enter your comment title...

Posted by Qiong on 30 Mar 2012 at 20:15 GMT

This paper used the mutations of gene TP53 of tumor protein P53 as their training data and used four machine learning methods to train mode to predict the functional impact of missense variants in BRCA1. It's good, but from the statistical point of view, it should be better if we can use mutations from more than on protein as the training data.

No competing interests declared.