BLASTp command line : filtering results by a minimum % of identity and % coverage

Question

[A_heath]

Hi all,

I use regularly BLASTp analysis and I would like to improve my command line so that only resulting hits with 60% of similarity and 60 % of coverage appear in the output file.

For the coverage, I use the following option that works great:

-qcov_hsp_perc 60

However, does anyone know what is the option to filter reads over 60% of similarity using blast directly?

I usually do it manually using an awk command but I would rather use blast to filter the resutls directly if possible.

Thank you very much for your help and advice!

[lieven.sterck]

well there is the -perc_identity option for blast.

However, do keep in mind that it is not trivial to get the numbers you would like directly from blast. One example: most of those values are provided on an HSP basis (so no full alignment or such) and might thus be very different when you compare two full proteins to each other

[A_heath]

Thank you so much for your reply. Indeed, that's exactly what I'm trying to avoid. Is there a way using blastp to have whole length protein comparisons? and not just HSP. I tried using this option but infortunately I have an error message : "unknown argument"

[GenoMax]

BLAST is a local aligner. If you need full length alignments then you may want to look at a global alignment program.

[lieven.sterck]

what GenoMax said ;-)

plus: I think it is possible that parameter only works for blastN and not blastP (I just checked and indeed, that parameter is not available for blastP. not surprisingly as %identity is not really applicable there)

[A_heath]

Thank you both for your much appreciated help!