Hi all,
I have searched LTQ-FT data against a decoy database using X!Tandem/k-
score. As for the xinteract command, I enable the accurate mass
binning and use the decoy hits to pin down the negative distribution.
My question is regarding the non-parametric model option. By enabling
it, I usually get more peptide identifications. But the truth is that
I don't quite understand what this non-parametric modeling is. I did
read several posters and papers from Nesvizhskii, Shteynberg, among
others, but I couldn't find a more detailed answer. Can anyone help a
non-statistician?!?!?!

One more question, I understand that most people use a 'relaxed'
precursor tolerance, say 2 Daltons, to search non-decoy database with
XTandem, and then use Peptide Prophet with accurate mass binning. What
about searches against decoy databases?? Do 'relaxed' tolerance work
better than more strict tolerances?? I have tried 2Da against 20 ppm
decoy searches both using ACC MASS BIN and haven't seen much
difference. Has anyone have a different experience??

I'd appreciate any advice.
Thanks,
Alex
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"spctools-discuss" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/spctools-discuss?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to