Blastp Against Human Proteome
2
2
Entering edit mode
13.0 years ago
Sophie ▴ 20

Hey guys,

I have a list of gi's for about 10000 protein sequences all from vertebrate species. I want to blast these sequences against non-redundant protein sequences only from humans.

I will run the blast on a local cluster.

Can anyone help me with the appropriate command line that I need to use?

Thanks,

blast human proteomics • 4.4k views
ADD COMMENT
1
Entering edit mode
13.0 years ago
Darked89 4.6k

You may also try http://www.uniprot.org/batch/ (this one has 10k entries limit)

For accession/GI conversion if needed use: http://www.ebi.ac.uk/Tools/picr/

ADD COMMENT
0
Entering edit mode

I tried the ID mapping at Uniprot, however, it managed to map only 60% of the gi's. I'll now give a try with the conversion tool at EBI that you mentioned.

ADD REPLY
0
Entering edit mode
13.0 years ago
Sophie ▴ 20

I broke the problem into the following parts.

1) Get the fasta sequences for all the gi's 2) Create human refseq database (formatdb) 3) blastp -d custom_database -i human_fasta

For the first, I would normally go to Batchz Entrez, submit the list of gi's, and then send the fasta sequences to file. However, this is not working for 10000 gi's. I reduced the number to as low as 500, but NCBI stil does not generate the fasta file and reports an error - "Your session has expired. Please repeat your search".

Any suggestions on how to retrieve this many protein sequences given their NCBI gi's?

ADD COMMENT
0
Entering edit mode

This could be a temporary problem with NCBI servers. I used to download tens of thousands ESTs (using Taxonomy).

ADD REPLY
0
Entering edit mode

Thanks for the reply, darked89. I have been using the NCBI servers, particularly the Batch Entrez service, for quite a while now and have never experienced such a problem.

I first thought that expired session might be due to submitting too many gi's, but the problem persisted even when I reduced the number to reasonable levels.

Let's hope that it is a temporary server problem.

Do you know any alternative way to retrieve the sequences?

ADD REPLY
0
Entering edit mode

You may try idfetch program (http://manpages.ubuntu.com/manpages/dapper/man1/idfetch.1.html) from NCBI tools, which does not impose any limits for number of GIs. On ubuntu, you can get it by installing ncbi-tools-bin package. Alternatively, you can try EUtilities, but I think there were also some limits on number of GIs per request there.

ADD REPLY

Login before adding your answer.

Traffic: 2701 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6