pfscanV3 producing no matches for certain sequences

sib-swiss / pftools3

A suite of tools to build and search generalized profiles

GNU General Public License v2.0

10 stars 7 forks source link

pfscanV3 producing no matches for certain sequences #14

Open gsn7 opened 3 years ago

gsn7 commented 3 years ago

I have run pfscanV3 on several sequences e.g., https://www.uniprot.org/uniprot/E8WI69.fasta and I am not getting any matches using 2020_05 prosite release. If I use pfscan v2.3 I get the same match as when I scan the sequence on https://prosite.expasy.org/scanprosite/. Is this expected and is https://prosite.expasy.org/scanprosite/ not yet using pfscanV3?

smoretti commented 3 years ago

Could you give us exact command lines you used for both pfscanV3 and pfscan on E8WI69.fasta ?

gsn7 commented 3 years ago

for pfscanV3: bin/prosite/pfscanV3 -f --matrix-only data/prosite/2020_05/prosite.dat input/E8WI69.fasta

for pfscan v2.3 using either the ps_scan.pl script: bin/prosite/ps_scan.pl -d data/prosite/2020_05/prosite.dat --pfscan bin/prosite/pfscan -m -s -o ipro input/E8WI69.fasta

or directly bin/prosite/pfscan -f -kyz input/E8WI69.fasta data/prosite/2020_05/prosite.dat

smoretti commented 3 years ago

Currently two cases with ScanProsite:

query with a sequence -> pftoolsV3 are not used yet, but pftools V2. pftoolsV3 will be used in ScanProsite soon.
query with a UniProt identifier -> matches pre-computed with pftoolsV3 are returned.

About your sequence: it contains a circular profile. There are few of them in Prosite. They are always issues with them. The goal is to replace them but we need to find something as good. They are not currently implemented in pftoolsV3. A warning should be returned about it but it does not work.