Protocol Online logo
Top : Forum Archives: : Bioinformatics and Biostatistics

anybody interested in batch BLAST?? - (Oct/08/2002 )

In some cases like phage display, people got hundreds of sequences and don't like BLAST them one by one. However, the NCBI web-based BLAST program doesn't provide a batch BLAST ability.

QBLAST is a convenient way to access NCBI servlet. Again, QBLAST dosn't support batch BLAST. Does anyone have an idea how to run batch BLAST through QBLAST, or any other way?

thanks

Jing

-jing-

I am not sure if the Network Entrez version can do batch blasts or not.
Take a look at http://www.ncbi.nlm.nih.gov/Entrez/Network...z.overview.html

Network Entrez is a TCP/IP-based client-server version of WWW Entrez. This means it makes a direct connection with the NCBI databases over the Internet to retrieve its data. The data comes in a binary form taking up less network bandwidth during transfer.

SAGE

-sage-

Thanks for your interest in this issue. My inderstand is, all the web-based application of NCBI BLAST, like WWW BLAST, Entrez BLAST, QBLAST are not eligible for batch BLAST, because the cgi on the NCBI server accept query as a parameter of "Put" in this format: &QUERY=query_seq
As you can see that you can't submit more than one query here. That's hy i need write an method to submit the batch queries.

The question is:
1. How can i merge the Put and Get command, so that i don't need to "Format" the BLAST output for every query? (see: http://www.ncbi.nlm.nih.gov/BLAST/Doc/urlapi.html)
2. Should the 2nd query be submited after output of the fisrt one? In that case, it will be too slow to finish all queries.

Jing

-jing-

hi Jing,

I am sorry I misunderstood your question.

You can do batch blast using Bioperl. It has a module called standaloneblast which can read sequences in FASTA format http://www.bioperl.org/Core/bptutorial.htm...BLAST_locally_S. Bioperl also provides blast report parser. To use Bioperl, you need to install it on your computer (either Windows or Linux) and you need to have some knowledge of Perl.

SAGE

-sage-

Hi, it's nice that i can discuss with you here. Actually i already wrote a perl module to parse the blast report containing multiple BLAST results from a batch BLAST. It works fine.

Moreover, the NCBI BLAST standalone is eligible for batch BLAST. Only that the Web-based apllication doesn't, with the reason as i mentioned. I need to generate an batch BLAST application and put it on website, so that all people can use it free. It's not realistic to download BLAST Standardalone, generate a interface and let people run batch BLAST on YOUR machine, because i don't have so many resource, and i don't want to maintain the system, update database regularly...

Tha's why i need to use NCBI resource to do this job. I belive QBLAST is a convenient choice because of its URL API. Currently, i am thinking about collecting all RID first and then call Get command, with all the default parameter for g=Get command of course. Do you have idea about batch BLAST in this way?

Jing

-jing-