I need to retrieve a large list of protein and their related nucleotide sequences. However, batch entrez removes old GI numbers, which in my case are more than half. How can I override this, or which tool can I use to replace the web service?
P.D. I been trying also idfetch, but when I put a protein GI with a flag for protein database it returns the entire genome of the associated organism, so no good results.
In case that you know programming, a simple tool with these functions would be no problem. Strategy:
1. Extract each GI from the list (.txt file)
2. For each GI, fetch the corresponding data from Genbank. If data is available, continue step 3, if not (the GI is removed), skip it, back to step 1.
3. Check the length of sequence (by the length of sequence itself or just simply fetch the desired information at row "Locus" in the GB file. If the length is quite bigger (longer) than expected (i.e. whole genome), do align with your reference sequence to trim it, then store your modified data.
In case that you don't know programing, ask a programmer to help you.
Good lucks,
QTK
Dear Luis,
You could map the old accessions using EBI's picr (http://www.ebi.ac.uk/Tools/picr/).
Cheers
Use efetch. And you'll need to pull just the coding region for your gene of interest. There might be a simpler way than this, but I know this works...
The base $URL would be "http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?". Loop through your list of gi numbers or NP/ZP_ numbers (deprecated ones are OK) and store as $accession. Append the following to your $URL: "db=protein&id=$accession&rettype=gb&retmode=text". In Perl if you 'use LWP::Simple', you can do $dump=get($URL); and write to a file. Or you can parse it on-the-fly.
Loop through the output and look for "coded_by". Parse out the nucleotide accession number and the start/stop numbers for the coding sequence ($nuclaccession:$start..$stop). It could also be "complement($nuclaccession:$start..$stop)". Build another URL using the same base URL as above, but append the following: "db=nucleotide&id=$nuclaccession&rettype=fasta&seq_start=$start&seq_stop=$stop". If it's the complement, then also append: "&strand=2" to your URL. Use the efetch URL to download the DNA coding sequence.
Just input your GI identifiers on www.proteinside.org in "cutsom analysis" section and run the analysis. You will get your corresponding proteins in results (with a lot of other informations).
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools