Is It Possible For A Gene To Have A Mutation At The Same Position In Both Cosmic And Dbsnp?

Question

[Jordan]

Hi,

I have been analyzing the gene - BRAF. I realized that this gene has a mutation which is present in both COSMIC and dbSNP.

For e.g.,

chr7 140453129
chr7 140481476

Both these single base mutations are present in COSMIC and dbSNP. In COSMIC these mutations are labeled as substitution_coding_silent mutation

My understanding is that dbSNP has common mutations and COSMIC has mutations which are specific only for cancer. When I do variant analysis, I remove the common mutations, here dbSNP mutations and keep only COSMIC database mutations. But if I do this step now, I will be missing these mutations which are present in COSMIC.

Does anyone know why this occurs?

[Sean Davis]

dbSNP is not a database of "common mutations". It now contains many sources of variation, some of which are not common. Also, there are MANY variants in dbSNP that are disease-associated, so I would not recommend using presence in dbSNP as a filter without a little extra work to pull out truly common variants.

[Matt Shirley]

Up-voted your question, though, as I'm glad you asked, and this seems to be a common misconception that many people have.

Answer 1

[Matt Shirley]

I believe that the answer to your question is that COSMIC is not "mutations which are specific only for cancer". COSMIC is a database of mutations which have been observed to occur in cancer biopsies and cell lines, in specific genes. These mutations are reported from the literature, and there is no guarantee that these mutations will not appear in the germline of your favorite cell.

If I understand your issue correctly, you are saying that you would like to filter your variants based on whether they are annotated in COSMIC or dbSNP. You say that if you filter out the variants in dbSNP first and then keep only COSMIC mutations, you miss the mutations present in both databases.

My solution for you is to go back to the unfiltered variants and first extract only those in COSMIC, without filtering out the dbSNP mutations, and avoiding this problem entirely.