3
$\begingroup$

I have a long list of gene variations. For example, Here are 4 of them:

CBL Q249E

TERT H412Y

SF3B1 R625H

EGFR L747_T751delinsP

The first term identifies the gene, the second identifies the variation.

I am a data scientist with no background in biology or medicine. I am looking for an online resource where I could download synonyms, descriptive data, and other relevant info for all the gene variants.

I have seen two sites where data is offered for some gene variants, but not for others. For example, the EGFR L747_T751delinsP variant is well documented at this and this site.

While CBL Q249E is documented at the ckb.jax site but not the reactome site. Other variants are documented at reactome but not ckb.jax, etc.

Plus, there appears to be no easily downloadable file containing this data. Before I go scraping what I need off of a hodgepodge of websites, I wanted to check with you smart people to see if there is a user-freindly resource out there that has this data in a downloadable format.

$\endgroup$
3
$\begingroup$

Seems your variants are all possibly harmful mutations to human genes, in that case ClinVar can be quite useful, it returns hits to all your variants except L747_T751delinsP. I have never used the ClinVar API so I'm afraid I can't vouch for how easy to use that is.

When ClinVar doesn't have the variant (or info about it) you need you could search EBI to see if other sites have information for this variant, when queried with L747_T751delinsP EBI actually shows us that it has been entered into Reactome twice. (welcome to the mess that is sequence annotation!).

This might be outside of your scope but if you have an interesting variant that you cannot find information on you could try to find the most similar (known) variant using BLAST.

$\endgroup$
  • $\begingroup$ Yes, these are all potentially harmful mutations. Although passenger genes with mutations could ultimately be beneficial (harmful to the cancer), if I understand correctly. Thanks for these great resources. Could it be that L747_T751delinsP has 2 entries at Reactome because it is a type of variation that can occur in different genes? $\endgroup$ – ben Aug 31 '17 at 18:00
  • $\begingroup$ @ben seems like the only difference between the two entries is that the protein of "p-6Y-EGFR" entry has a bunch of phosphorylated tyrosine residues. Basically it seems the two entries cover the same gene, same variant in this gene and same protein sequence but the p-6Y entry has extra modifications to this protein. This probably does indicate difference in function but this should give you an idea of how confusing, and difficult to automate, this can get. $\endgroup$ – Koen vd H Aug 31 '17 at 19:42
1
$\begingroup$

You can use the Table Browser.

For the variants, in the group field select variation, in the track field, select any database you are interested in, finally insert the genome position of your gene of interest and you will be able to download the relative data.

For the synonyms, descriptive data, and other relevant info for all the gene variants, in the group field select Phenotype and Literature and from the track field, select any database you are interested in. You may need to download data from multiple databases to get all the info you want, but the whole procedure is quite fast and you can do it directly from the browser.

$\endgroup$
  • $\begingroup$ Thanks. I do have location info for each of the genes in my data (which I got from a website--can't remember which off the top of my head right now), so I can fill in the important fields at Table Browser. Problem now is that the names of the variations given at this site do not match any in my list of variations. They all start with "rs" followed by a number, whereas you can see in the original question above the variations in my data follow a different nomenclature. Any idea how to translate between the two? $\endgroup$ – ben Aug 30 '17 at 23:37
  • $\begingroup$ @ben It's not trivial to find the correct ID, different databases use a different one. The rs suffix is from the dbSNP database. The mutation you are looking for EGFR L747_T751>P is an indel, you can find it on the Cosmic database $\endgroup$ – alec_djinn Aug 31 '17 at 9:51
1
$\begingroup$

I think the correct nomenclature is EGFR L747_T751delinsP, based on http://varnomen.hgvs.org/recommendations/protein/variant/delins/

You can also use a translator tool like TransVar

@smock2020 and @JAX_CKB

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.