Extract multiple sequences on ncbi by gene symbol

Giorgio C · 01-24-2013, 07:14 AM

Hi all,

I have a lot of gene symbols belonging to the same organism and I need to extract the corresponding sequences on ncbi.

The gene symbol are all like: NMB0001, NMB0010, NMB0015...

Is there any automated way (scripts, tools,..) in order to obtain the fasta sequences?

Thanks in advance,
Giorgio

Richard Finney · 01-24-2013, 08:43 AM

Download gene_info and gene2accession from NCBI.

run these :
"wget -nc ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene_info.gz"
"wget -nc ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene2accession.gz"
Run gunzip to get these two files uncompressed and readable.

Get your genename(s) converted to entrez gene ids (second column), make sure taxonomy id is 122586 (. in your case).
Convert the gene names to GIs (gene info identifiers) or accessions.

Like this ...

bash-3.00# grep NMB0001 gene_info
122586 902103 NMB0001 NMB0001 - - - - acetyltransferase protein-coding - - - - 20120121

bash-3.00# grep "122586.902103" gene2accession
122586 902103 - - - AAF40480.1 7225226 AE002098.2 66731897 - - ? -
122586 902103 PROVISIONAL - - NP_273067.1 15675949 NC_003112.2 77358697 6 497 - -

Cut out the the accession(s) you want, either RNA or Protein. Put them into a flat text file as a list.
You'll want to program using bash/perl/c/command line utilities, whatever to automate the two basic commands above. Be careful.

There's supposedly "batch entrez" : http://www.ncbi.nlm.nih.gov/sites/batchentrez

But I've found it to be unreliable. It drops some accessions. Sad but true. You're welcome to try it.

The absolute best utility for batch downloading is "idfetch" : http://man.cx/idfetch%281%29 . It's command line, takes parameters, you can give it a file of your accessions or GIs and crank it up. ....downloading ... downloading ... downloading ... wahlah ... your file appears ...just like you asked for. I keep an old copy from an old NCBI C toolkit around and ... apparently it's still available in ncbi-tools-bin package ...

I run this on Linux/Unix command line ...
sudo apt-get install ncbi-tools-bin
and idfetch is there !!! So you can install the ncbi-tools-bin package and get it.
___
sudo apt-get install ncbi-tools-bin

$ man idfetch | head
IDFETCH(1) NCBI Tools User's Manual IDFETCH(1)

NAME
idfetch - retrieve biological data from the NCBI ID1 server

SYNOPSIS
idfetch [-] [-F str] [-G filename] [-Q filename] [-c N] [-d str] [-e N]
[-f str] [-g N] [-i N] [-l filename] [-n] [-o filename] [-q str] [-s str]
....

Giorgio C · 01-24-2013, 09:49 AM

Thank you so much for your time and suggestions. Actually I tried 'batch entrez' and as you reported, it's not so good; so you're right about that.

I really like your last option and I'm gonna try that.

Thanks again.

Cheers,
Giorgio

dexterslab · 01-24-2013, 06:49 PM

hi,
you might also wanna look into the BiomaRt package for R.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
how to convert UCSC gene names to gene symbol	gigigou	Bioinformatics	5	05-19-2015 05:23 AM
A script for Cuffdiff: extract differential expression gene sequences from genome.fa	foxriverlin	Bioinformatics	17	04-11-2013 03:46 AM
How do I extract partial sequence data (Fasta) from multiple hits in NCBI nucleotide?	andtill	Bioinformatics	1	11-09-2012 01:49 PM
Extract intergenics regions from ncbi	RTK45	Bioinformatics	1	03-21-2012 02:24 PM
CUFF_id -> Gene symbol tool?	Inti	Bioinformatics	0	03-18-2011 03:58 AM

01-24-2013, 07:14 AM	#1
Giorgio C Member Location: ITALY Join Date: Oct 2010 Posts: 89	Extract multiple sequences on ncbi by gene symbol Hi all, I have a lot of gene symbols belonging to the same organism and I need to extract the corresponding sequences on ncbi. The gene symbol are all like: NMB0001, NMB0010, NMB0015... Is there any automated way (scripts, tools,..) in order to obtain the fasta sequences? Thanks in advance, Giorgio

01-24-2013, 08:43 AM	#2
Richard Finney Senior Member Location: bethesda Join Date: Feb 2009 Posts: 700	Download gene_info and gene2accession from NCBI. run these : "wget -nc ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene_info.gz" "wget -nc ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/gene2accession.gz" Run gunzip to get these two files uncompressed and readable. Get your genename(s) converted to entrez gene ids (second column), make sure taxonomy id is 122586 (. in your case). Convert the gene names to GIs (gene info identifiers) or accessions. Like this ... bash-3.00# grep NMB0001 gene_info 122586 902103 NMB0001 NMB0001 - - - - acetyltransferase protein-coding - - - - 20120121 bash-3.00# grep "122586.902103" gene2accession 122586 902103 - - - AAF40480.1 7225226 AE002098.2 66731897 - - ? - 122586 902103 PROVISIONAL - - NP_273067.1 15675949 NC_003112.2 77358697 6 497 - - Cut out the the accession(s) you want, either RNA or Protein. Put them into a flat text file as a list. You'll want to program using bash/perl/c/command line utilities, whatever to automate the two basic commands above. Be careful. There's supposedly "batch entrez" : http://www.ncbi.nlm.nih.gov/sites/batchentrez But I've found it to be unreliable. It drops some accessions. Sad but true. You're welcome to try it. The absolute best utility for batch downloading is "idfetch" : http://man.cx/idfetch%281%29 . It's command line, takes parameters, you can give it a file of your accessions or GIs and crank it up. ....downloading ... downloading ... downloading ... wahlah ... your file appears ...just like you asked for. I keep an old copy from an old NCBI C toolkit around and ... apparently it's still available in ncbi-tools-bin package ... I run this on Linux/Unix command line ... sudo apt-get install ncbi-tools-bin and idfetch is there !!! So you can install the ncbi-tools-bin package and get it. ___ sudo apt-get install ncbi-tools-bin $ man idfetch \| head IDFETCH(1) NCBI Tools User's Manual IDFETCH(1) NAME idfetch - retrieve biological data from the NCBI ID1 server SYNOPSIS idfetch [-] [-F str] [-G filename] [-Q filename] [-c N] [-d str] [-e N] [-f str] [-g N] [-i N] [-l filename] [-n] [-o filename] [-q str] [-s str] .... Last edited by Richard Finney; 01-24-2013 at 08:56 AM.

01-24-2013, 09:49 AM	#3
Giorgio C Member Location: ITALY Join Date: Oct 2010 Posts: 89	Thank you so much for your time and suggestions. Actually I tried 'batch entrez' and as you reported, it's not so good; so you're right about that. I really like your last option and I'm gonna try that. Thanks again. Cheers, Giorgio

01-24-2013, 06:49 PM	#4
dexterslab Junior Member Location: eastern part of this planet Join Date: Jan 2013 Posts: 5	hi, you might also wanna look into the BiomaRt package for R.