download protein sequences from NCBI

Question

[guillaume.rbt]

Hi all,

I would like to download all protein sequences from one species on NCBI:

https://www.ncbi.nlm.nih.gov/protein?linkname=bioproject_protein&from_uid=261773

This is maybe trivial, but is there a way to download all sequences concatenated in only one fasta?

Thanks a lot,

Guillaume

[Prasad]

there is a send to option through which you can download all the sequences. After just remove the fasta headers to make a single fasta

awk 'BEGIN{a=0}{if($0~/^>/){if(a==0){print}a++;}else{print}}' input.fasta >out.fasta

[guillaume.rbt]

thanks for the response, how do you use the send to option? Is this on the console or on the website?

Answer 1

[Pierre Lindenbaum]

using the ncbi interface you can just click on "Send to > File"

or using eutils:

curl "http://www.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=bioproject&id=261773&linkname=bioproject_protein"  | xmllint --xpath '//LinkSetDb' - | xmllint --format - | grep "<Id>" | cut -d '>' -f 2 | cut -d '<' -f 1 | while read L; do curl "http://www.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=protein&id=${L}&rettype=fasta" ; done

>gi|821074095|gb|KKY28990.1| putative uracil phosphoribosyltransferase [Diplodia seriata]
MFVHASGPESIKFKHLQGQVQVLLVDSVINSGATILDFVEAIREINPGIRIVVVAGTVQAQCISPNNPFY
KTLAQHGDISLVALRSSETKFTGSGGTDTGNRLFNTTHLL

>gi|821074094|gb|KKY28989.1| putative integral membrane protein [Diplodia seriata]
MPQYFPWPYSVDPLPEDLRRGLWPVGIFALMSTVATLALLCWITYRLVSWRKHYRSYVGYNQYVLLIYNL
LLADLQQSISFLISFHWIHTDSMLAPSPACFGQAWLVQIGDISSGMFVLAIALHTFFSVVKGRQIPFRAF
LIGTIVIWALALLLTVLGPALHGSDYFTAAGAWCWASDKYETERLWLHYLWIFIIEFGTVIIYALIFIYL
RKQLVSIASAHQHSTQNKVSQAARYMVLYPLTYVLLTLPLAAGRMATMTGQTLPIAYYCAAGSMMTSCGW
VDAALYALTRRVLVSNEIDQPQGGAGKGASSSGGRTGYGGHGSSHTATGWDIASFSDRKGGMGADHSVTI
TGGLDARGSNFIDMDELSKGGVHHHATERVGRPKHKGSSTPSTQGLTRARSSSTSARESTPRGSTDSILA
GLGGVRAETKVEIRVEPANGFMLPGEGSGSNGSSGMSTPNGRTVEVVGNSHAMRPRSGSPY

Answer 2

[tlorin]

Here is a well-explained tutorial for your problem :)

[Sej Modha]

The link you provided doesn't seem to work.

[tlorin]

This should be better now, thanks!

[guillaume.rbt]

thank you all for your help! works fine

Answer 3

[Sej Modha]

NCBI Unix e-utils version of the @pierre's solution

esearch -db bioproject -query 261773|elink -target protein |efetch -format fasta

Answer 4

[Kevin Blighe]

Another solution(s) implemented as a Python script:

• A: How to download all sequences of a list of proteins for a particular organism
• A: Need help to retrive sequences

Kevin