Which one is the best tool for predicting both gain-of-function and loss-of-function mutations if you have to select only one server from PolyPhen and SIFT?
Generally, We usually consider both in a complementary approach, polyphen-2 may have a wider approach for accuracy improvement, but you may expect some mis-alignments in the prediction results so it's recommended to use both & compare results. you may also consider integrative tools that include both.
Hi. Generally, the is no consensus for choosing one tool over the other and reporting the result of only one prediction tool would not make reviewers and editors happy. You can combine the result of several tools (SIFT, Polyphen-2 etc) and choose a cutoff of, let's say, 3 out of 5 tools must predict damaging effect, or, use a more inclusive tool such as CADD. Here, CADD > 20 could be a reasonable threshold for predicted pathogenicity.
For loss/gain of function mutations I would also consider the gene's genic intolerance against lof mutations, i.e. the tolerance (expected vs observed lof rate) for lof mutations, ExAC LoF FDR here: http://genic-intolerance.org/index.jsp. Some genes are particular deficient of lof mutations and these could be of special interest. //Hans
Hi everyone, I have a similar question, but different regarding what can one use to predict loss or gain of function of a particular variant?? Or what should one check to consider loss or gain of function from a particular variant?