Reasons and reference for protein/peptide threshold value

[AnaP]

I need an information from you, if you can help. A long time ago, an
ex-member of your team suggested that for Peaks Protein ID I should
consider protein/peptide Peaks scores of >49. He tried to explain why
this is, but I need further information on this. For a publication I
am preparing, I need to indicate the threshold score I will use and
also reference it, justify why would it be scores >49, for example.
Have you in Bioinformatics Solutions ever published (or in a
collaboration or other publication) details on the Peaks Protein ID
scoring? I am more interested on the peptide score threshold. What
score would be the equivalent of the confident Mowse score (P>0.95)
threshold?

Thank you very much for your help.

Kind regards,

AP

[Lynette]

Dear AP,

Which version of Peaks are you using? Peaks 4.5 peptide scores are
based on benchmark statistics so it would be dependent on the
instruments and the quality of your data. We basically did statistics
and found that PEAKS scores for protein identifications are either
much larger than 50, or much lower than 50. There're only very few
proteins at the threshold boundary. In addition, the score-larger-
than-50 proteins usually have good qualities (e.g. with multiple
peptide support). However we maybe able to help you in justifying the
score to use for protein ID in your publication. In Peaks 5.0 we would
recommend you use our decoy database validation function to determine
the false positive rate to chose an optimal threshold.

Cheers,
---------------------------
Lynette Lau, MSc
Bioinformatics Solutions Inc.
(519) 885-8288 Ext 14
http://www.bioinfor.com