Annotation databases for somatic mutations

Question

[bstrs]

It would be great to have a list of somatic mutations databases that are out there (that is one looks into that database and sees if a given mutation is somatic or not and what else is known about it if the mutation is somatic).

I am aware of these:

• COSMIC: https://cancer.sanger.ac.uk/cosmic
• OncoKB: https://www.oncokb.org/
• ClinVar: https://www.ncbi.nlm.nih.gov/clinvar/
• CIViC: https://civicdb.org/home
• PMKB: https://pmkb.weill.cornell.edu/
• DoCM: http://docm.info/

Any other databases out there?

[ATpoint]

A database can never tell you whether a given mutation in your specimen is somatic or not. Somatic means that it was not inherited by (or from) the parents but occured de novo in that specimen, nothing more, nothing less. Databases might tell you allele frequencies in certain populations and known impact of variants in certain cellular contexts but even the super rarest of mutations can still be due to germline inheritance. It is then on you to find out whether that mutation is somatic or not by comparing the observed mutations with other cell types from that donor.

[bstrs]

@ ATpoint

Actually, a database can tell if a mutation is known to be somatic or not! For example, Ensembl database does that. See screenshot here: https://ibb.co/n6m6FtT

Some databases, might call them oncogenic, pathogenic or whatever. Of course no one expects that info from databases to be 100% accurate either. And having a somatic mutation database that shows that a mutation has been "found to be somatic mutation in previous cases" helps a lot.

So one can look to literature and previous reported cases and still figure out if a given mutation is likely somatic or not.

[ATpoint]

So, say a father and a mother are both homozygous for such a mutation, so will be all offspring. Makes the variant germline in the offspring, regardless of what any database says. But I see your point, can be informative to know whether a mutation is somatic in the majority of cases, you are right.

[bstrs]

Of course, you have a point with your example. But here is a concrete example.

For example, the BCR-ABL fusion (or TMPRSS2-ERG fusion) is a somatic mutation. It is known in the scientific literature as such. Nobody just inherits it from the parents. So if ones finds it in a NGS data and looks into the literature will see that this somatic mutation is known only and only as somatic mutation and has never been found as germline mutation.