"Default protein report"

[Wannes Dermauw]

Hi 

I generated a "Default Protein report" with Peptideshaker (for an LC-MS/MS analysis of a control sample), but I am a bit confused regarding the confidence levels of the proteins that are included in this report (see example attached).

These are all proteins (and mainly CRAP proteins) with very low confidence levels (below 10%), yet they are retained in the "Default Protein Report". Should I consider these proteins as reliably  "identified" or rather ignore these identifications because of their low confidence. In case of the latter, why are these proteins then included in the "Default Protein Report". (FYI, default FDR settings were used in Peptideshaker)

FYI, for other (treatment) samples "Default Protein Reports" contain proteins with high confidence levels (=100%).

Any suggestions are more than welcome.

Kind regards

Wannes

[Harald Barsnes]

Hi Wannes,

Sorry for the very late answer! Seems like there was an error in the notification system for the Google group and I did not see your message until now.

The confidence for the proteins in question indeed seems to be a bit low. It is however hard to judge the confidence for a couple of isolated proteins without having the complete distribution of confidence values. I would recommend having a closer look at the Validation tab in PeptideShaker to see if there might be something strange with the validation of this particular dataset. Perhaps there are a low number of high-scoring target hits? Or a low number of high-scoring hits overall?

It is usually also recommended to combine the confidence value with the information in the Validation column, as this may provide additional information about whether a given hit ought to be trusted or not.

Best regards,

Harald