Include description of correct features (i.e. -D parameter) in Percolator

[MiguelCos]

Hello,

I am curious about the parameters used in the web version of Prosit for the FDR calculation.

For your publication, I understand the pipeline sets up the flags -Y and -U... I was wondering why you did not also set up the -D flag to "Include description of correct features, i.e. features describing the difference between the observed and predicted isoelectric point, retention time and precursor mass"?

Wouldn't it improve the FDR calculation since Prosit is also using RT information as part of the prediction of fragment spectra and peptide to spectrum matching?

Maybe I am missing something regarding the way the algorithm works and I would love to better understand it.

Best wishes,
Miguel

[tkschmidt]

with the next published version of the online prosit (coming soon, we are doing the last dev. cycle) we are using predicted and fitted iRT from Prosit.

When we evaluated the iRT feature for the publication, it was a quite strong feature but when using all the spectral feature it lost most of its power :/ Still, for bigger searches its helpful and we will include it within the next week(s).

P.s. wasn't aware about the isoelectrical point prediction. is it a powerful feature?

[MiguelCos]

Hello Tobias, thanks for your answer.

Then I understand that the Percolator parameters in the web pipeline do not include the -D flag and don't consider these RT features. I have not tested this isoelectric point feature in the model for FDR calculation thoroughly, but it would be interesting to have this info and the RT in the .tab output files after prosit to be able to tweak those parameters in percolator after prosit.