Is there a way to predict gain-of-function or gain-of-function based on the amino acid change?
I've been using ANNOVAR for variant functional annotation. It has many databases to add commonly used info. These include resources like SIFT or PolyPhen, but those just give pathogenicity scores. Thus, this makes me think this may not be so trivial.
Because it's not trivial. :)
Most loss of function events will be non-synonymous mutations in exonic regions which either significantly truncate or frameshift the resulting protein. You don't really know whether it's loss of function without doing validations or looking at protein models or crystallography to assess structural features. SIFT / ClinVar scores etc are usually good indicators of these types of events. Same really goes for gain of function. What you're likely looking for are mutations with high significance annotations (i.e., likely to result in significant alterations to protein composition, e.g., stop gain event, early frameshifts, etc) and high ClinVar scores for any clinical diagnostics. If you're looking for novel SNP / indels associated with a phenotype, you could do GWAS then validate hits invitro / invivo. But that's pretty nuanced in it's own right. Not something you just pick up in a weekend.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Damn right.
Except for very obvious cases (deletion of a crucial exon, mutation of crucial cysteine) it's rather hard/impossible to predict the functional consequence of e.g. a random missense variant.
I would think that mutation pathogenicity estimates are difficult as well. Not to imply they are easy, but at least there are many attempts (SIFT, PolyPhen, MutationTaster, MutationAssessor, FATHMM, PROVEAN, MetaSVM, MetaLR, VEST, M-CAP, CADD, GERP++, DANN, etc.).