NCBI Blast Automation

[andy28_sd]

Hi! I am new at programming (study back in college) and biotech field. I am task to write a program that will do a sequence blast against NCBI database and return top 10 results.

1) I was thinking about using NCBI Blast WSDL and when someone enter the sequence into Oracle DB. It will go to NCBI and return the sequence result.

2) Java NCBI blast client, export all the sequence at night and do batch blast against NCBI blast.

3) Any other ways someone think it is better route.

So, if anyone know which way is better or know the script to do it, please let me know.

Thank you!!!
Andy

[WhatsOEver]

NCBI offers an example perl script to do a batch blast http://www.ncbi.nlm.nih.gov/IEB/Tool...t/web_blast.pl

And if you're going to use it, remember this part:

Please do not submit or retrieve more than one request every two seconds.

Cause if you are too greedy, they may block you for some time

[arundurvasula]

Hi, NCBI offers blast+ for this: http://blast.ncbi.nlm.nih.gov/Blast...._TYPE=Download
which is what I use. You can then blast a file of sequences using a command like this:

blastx -db nr -query contigs.fa -evalue 10 -matrix 'BLOSUM62' -word_size 3 -gapopen 11 -gapextend 1 -max_target_seqs 3 -outfmt "10 std stitle" -out ../results/blast/blast-002.csv -remote

[GenoMax]

If you expect to do this regularly (and with lots of sequences) then setting up local blast (along with regular download of blast indexes) may the thing to do.