Uniprot Refseq References Not Included In Refseq Complete Release

Question

[Pablo Pareja]

Hi all!

I recently came across this situation when parsing the whole Uniprot Sprot and Trembl xml files.

There are some cases where references to RefSeq in Uniprot entries are not included in RefSeq complete release. After looking for some of the problematic references (randomly selected), it looks like they are all provisional records; but then, my question is:

Why are they included as a reference?

Does any of you have more information about this?

Here you have some of the problematic references: (I could provide the full list if anyone may find it useful)

Sprot references:

Q8D8G9 , NC_004459.3

P25445 , NM_152871.2

P09470 , NM_009598.2

Trembl references:

Q0VAN4 , NM_181435.5

Q8DD16 , NC_004459.3

Q8D4H5 , NC_004460.2

Cheers,

Pablo Pareja

Answer 1

[Pablo Pareja]

Hi all,

Here you have the answer I got from uniprot:

As per protocols we setup with NCBI, we are mapping all RefSeq records but the ZP_ entries (translated from whole genome shotgun sequences).

Our goal in mappings is to have as much as possible coverage to RefSeq. Currently Provisional, Predicted, Inferred or Model based entries account for > 95% RefSeq protein. So we do not want to miss such a large chunk in UniProtKB mappings.

The RefSeq cross-references to UniProtKB are not necessarily reflected in RefSeq records. Even if they did, given these two databases have different release cycles; the UniProtKB is released every 4 weeks and Refseq complete every ~2 months (with daily incremental changes), the cross-references in actual records will be out-of sync. On the other hand, as you have identified, the NCBI ensures the UniProtKB cross-references (as provided by UniProt consortium) on the web site are up-to-date. Further information can be inquired using NCBI helpdesk at [e-mail] at [e-mail].

Answer 2

[Jerven]

Hi Pablo,

For this specific type of questions I advice you to contact uniprot directly at [e-mail]. Then the question will end up on the desk of the people doing the linking and they should be able to explain why and how.

Regards, Jerven

[Pablo Pareja]

Hi jerven, I already did that, however I still think it's also a good idea asking it here. That way, much more people can benefit from the answer as well as provide the solution in case they already ran into similar situations. Cheers,

Pablo

[Jerven]

Of course feel free to add the reply to the page and mark it as answered.

[Jerven]

Of course feel free to add the answer they give to this page yourself. And mark it as an correct answer. Cheers.

[Pablo Pareja]

Yeah, I'll do that as soon as I get the answer. Cheers.