Minor PeptideProphetParser error with X!Tandem data with TPP 4.1.1

[bowe...@gmail.com]

I've only run some X!Tandem data (kscore) through the new version and
I encountered the error:
WARNING: Mixture model quality test failed for charge (1+). Use
xinteract option -OF (PeptideProphetParser option FORCEDISTR) to
override.

I added the FORCEDISTR option to PeptideProphet parser and the error
goes away as it should. However, on another dataset I am getting the
following error with the FORSEDISTR option enabled:
WARNING: Mixture model quality test failed for charge (5+). Use
xinteract option -OF (PeptideProphetParser option FORCEDISTR) to
override.


Here is the full output when PeptideProphetParser runs:
Not using ntt model
Forcing output of mixture model
(X! Tandem (k-score))
init with X! Tandem (k-score) trypsin
MS Instrument info: Manufacturer: Thermo Finnigan, Model: LTQ XL,
Ionization: NSI, Analyzer: ITMS, Detector: unknown

PeptideProphet (TPP v4.1 JETSTREAM rev 1, Build 200811211143
(linux)) AKeller@ISB
read in 658 1+, 642 2+, 468 3+, 163 4+, and 45 5+ spectra
Initialising statistical models ...
Iterations: .........10.........20........
WARNING: Mixture model quality test failed for charge (5+). Use
xinteract option -OF (PeptideProphetParser option FORCEDISTR) to
override.
model complete after 29 iterations

Greg

[Natalie Tasman]

(I don't think my reply went through this afternoon, so reposting:)

Hi Greg,

PeptideProphet will fail the model if there is too much overlap between postive and negative distributions.  We should update the warning message and actually not recommend using this option.  FORCEDISTR is really only to be used by developers as an aid towards understanding what would happen if PeptideProphet's automatic checks were bypassed.

However, given that it should "work" in the sense of bypassing the automatic check, please upload a file and David S. will take a look at it.

Thanks,

Natalie