Obtaining nucleotide gene sequence of bacteria using eutils

[jpearl01]

Hi All,

I'm trying to write a quick program to download a bunch of genes' dna from bacteria. I don't have ids for the genes, just their protein names (like lacz or what-have-you). So use esearch to look for say the lacz bacteria gene from the gene database:

http://eutils.ncbi.nlm.nih.gov/entre...p+AND+bacteria[filter]

then, I grab the first id in the list that is returned, and convert it from a gene id to a nucleotide using elink

http://eutils.ncbi.nlm.nih.gov/entrez/eutils/elink.fcgi?dbfrom=gene&db=nucleotide&id=945006

Which gives back a list of different ids. Most of the ids (using efetch to get the sequence) generally give back the entire genomic sequence of the organism it was found in. *sometimes* one of the id's gives back the actual gene nucleotide data, but not always. For example the first two ids from the above elink result give whole genomic sequence:

http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=nuccore&id=556503834&rettype=fasta

The third id gives the gene sequence:

http://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=nuccore&id=342209847&rettype=fasta

So, what gives? Is there a way to tell which ones are whole genome sequences, and which are gene sequences? Maybe an elink parameter?

~josh

[TiborNagy]

Try to download the sequences in GeneBank format and parse the descriptions.

[jpearl01]

Well, that is a possibility. However, I would really like it if I didn't have any intermediary steps. What is frustrating is that if I do the same search, but then link to the *protein* database, I get the correct AA sequence almost every time. I'm not sure why I can do that for the protein sequence but not for the nucleotide. Really, I'm wondering if I'm doing something wrong.

[gringer]

Note that NCBI has now released Entrez Direct, which has pipeable functionality for downloading via Entrez. There is likely a specific query that can be fed into Entrez direct for the bacterial genomes.