PeptideMapper does not find occurences with more than one 'X'

[cbielow]

Hi,

I'm doing some testing for PeptideMapper and found a restriction, namely it seems to consider at most one 'X' per match.

Here is a small demo:
Given proteins PEPTIDEMAPPER_TEST.fasta:

sp|GGGG|...
GGGG
sp|GGGGX|...
GGGGX
sp|GGGGXX|...
GGGGXX
sp|GGGGAA|...
GGGGAA

and the peptides PEPTIDEMAPPER_TEST_peps.csv:

GGGGAA
GGGGA
GGGG
GGGGXX

, PeptideMapper will report this:

GGGGAA,GGGGAA,1
GGGGA,GGGGAA,1
GGGG,GGGGXX,1
GGGX,GGGGXX,2
GGGX,GGGGX,2
GGGG,GGGGAA,1
GGGGX,GGGGX,1
GGGG,GGGGX,1
GGGGX,GGGGXX,1
GGGG,GGGG,1

thus, peptide GGGGAA will not match to protein GGGGXX; it seems matches are only allowed to contain 1 'X at most.
Also peptide GGGGXX will not match to protein GGGGXX (which would be an exact match).

Is this intended behaviour?

Invocation was:
java -cp $FM_jar com.compomics.cli.peptide_mapper.PeptideMapperCLI -p ../databases/PEPTIDEMAPPER_TEST.fasta ../databases/PEPTIDEMAPPER_TEST_peps.csv peptidemapper_result_minitest.csv -c 1

[mvaudel]

Hi Chris!

It has been a long time, I hope all is well! If you have a minute, I would love to have news from you and the team in Berlin :-)

Regarding this issue, by default PeptideMapper returns matches that have max 25% of 'Xs' in the sequence. My interpretation is that GGGX (25%) is accepted but GGGGXX (33%) is not. You can relax this threshold using a parameters file (UtilitiesParameterFile) that can be created using IdentificationParametersCLI, and there you can set Sequence Matching parameters.

I am sorry that this is convoluted and not clear from the documentation. If it is something that can help, I can add sequence matching options directly in the PeptideMapper command line.

Kind regards,

Marc