Criteria for protein identification/quantification

[MR]

Hi all
I`m a new user of Peaks and was hoping someone could give me an
indication of the cut-off points they use for assigning reliable
protein IDs. ie is there a threshold %score that can be used, a
certain number of matched peptides and coverages etc?

Also, is there a way of estimating the relative abundance proteins in
a mixture using MS/MS data? I`ve read in a paper that the "spec
count" (the total number of times all of the unique peptides of a
given protein were identified) is a quantitative measure of the
abundance of a protein in a mixed sample...
Is there anything comparable in peaks?

Any suggestions would be greatly appreciated.

Cheers

MR

[iainiai]

Hi MR,
I like to use a cut-off of "two high-scoring peptides". And you can
set a filter for this. Add a protein queries filter >1 and a peptide
score filter >50. At the peptide level, 50% is a good score cut off
for PEAKS.

For the quantitation, I don't believe spec count to be a reliable
measure, and I'd rather take the sum of peak intensities for a protein
as an estimate of its abundance in relation to another protein.
There's no indication of this in PEAKS Studio, but the team has worked
on protein abundance using ICAT and SILAC type labelling.

-IainR