This is a test version of Biostars. For the public version, visit https://www.biostars.org.
short protein BLAST

Dear All

I am performing BLAST. The database is a collection of short protein sequences (~40 aa in length, short but not short very much). The query file contains short protein sequences as well (~40 aa). I found BLASTp E-value cutoff of 1e-5 works badly. What's your suggestions on doing this BLAST? Thanks a lot!

blast
• 3,635 views • link
updated by Ram Ram • written by biolab biolab

"Performing BLAST" is a very bad descriptor. What is your goal doing this? Are you looking for similar sequences? Proteins with a specific domain? Proteins which has not been observed before?

And what do you mean by "short, but not short very much"?

• 1 views • link
updated by Ram Ram • written by David Westergaard David Westergaard

1 answer

What do you mean by it works badly... too much sequence are retrieved ? Proteins are not really similar ? ... ?

Your e-value depends of the size of your query (m) and the size of your database (n).

See here: http://www.ncbi.nlm.nih.gov/BLAST/tutorial/Altschul-1.html

So, you can try by cheating on your database size (you can define with a parameter), or use the bit-score instead of the e-value. You can also decrease your cutoff.

I invite you to consult also this document: http://homepages.ulb.ac.be/~dgonze/TEACHING/stat_scores.pdf.

• 1 views • link
updated by Ram Ram • written by Juke34 Juke34

Thank you very much for your suggestions.

• 0 views • link
written by biolab biolab

Log in to answer this question.

More posts like this