Problem with web Blast SRA database

Question

[QVH]

Hello,

I want to blast some sequences on dozens of huge SRA datasets. And, in order to develop a pipeline, I tried to send different requests from my computer on python. Everything is working fine with the standard parameters (database = nr, for example), but I don't manage to make it work when I want to use a SRA dataset as a Database (DRX029854, for example, with the BLAST_SPEC:SRA parameter).

fasta_file = 'BlastList.fasta'.strip()

database_num = str('DRX029854')

url = ('http://blast.ncbi.nlm.nih.gov/Blast.cgi')

args = {'CMD':'Put','DATABASE':database_num,'PROGRAM':'tblastn','BLAST_SPEC':'SRA',
            'FORMAT_TYPE':'XML','MAX_NUM_SEQ':'20000'
                        ,'WORD_SIZE':'6','FILTER':'F'}

req = requests.post(url,params=args,files={'QUERY': open(fasta_file, 'rb')})

Does someone know which API or any kind of specific parameter I have to use to make it work with SRA databases?

Thanks

[frcamacho]

Maybe take a look at this code: https://github.com/Kingsford-Group/sbtappendix/blob/master/srablast/srablast.py

[QVH]

It's the same kind of code I'm using, and unfortunately, it's not working. I get this error :

Informational Message: No alias or index file found for nucleotide database [DRX029854] in search path [/export/home/splitd/blastdb/blast0:/blast/db/disk.blast/blast1105::]

I think the problem could be due to the way the database is called.

[GenoMax]

Appears that the script is looking at the local file paths rather than web for the database.

Are you able to get this to work for some other acc #? Perhaps your local firewall settings are preventing you from going out to NCBI.

Also be aware that at the time I write this NCBI is testing https only access to their site. So you may want to replace that http with https (https://blast.ncbi.nlm.nih.gov/Blast.cgi). just in case.

[QVH]

It's perfectly working when I try to blast something on a standard database (e.g. 'nr'), for exemple:

fasta_file = 'test-tblastn.fa'.strip()
database_num = str('DRX029854')
url = ('https://blast.ncbi.nlm.nih.gov/Blast.cgi')
args = {'CMD':'Put','DATABASE':'nr','PROGRAM':'tblastn',
            'FORMAT_TYPE':'XML',
                        'WORD_SIZE':'6','FILTER':'F'}

req = requests.post(url,params=args,files={'QUERY': open(fasta_file, 'rb')})

So, I really think the problem is due to the way I call the SRA database.

[GenoMax]

Do you know for sure that NCBI allows remote access to SRA database for blast?

[QVH]

I contacted NCBI in order to know whether this function has been removed.

[QVH]

Looks like it's not "officially supported" by NCBI, they recommend cloud implementation with Amazon.