Questions about spectral library generation different search engines and shared peptides

[Nikita Boeren]

Dear TPP team and users,

I am a graduate student and I am trying to figure out a few things regarding spectral library generation using TPP for SWATH analysis. My questions do not involve direct issues with the use of TPP, but more about the results.

My workflow is: Comet or X!Tandem search, PeptideProphet, iProphet, Mayu (1% protein FDR), SpectraST and SpectraST2TSV. 

I have used both Comet and X!Tandem search engines to create two libraries. Also, I combined those two search outputs (via iProphet) to create another library. In contrast of what I expected, my final list of proteins in my combined library is less than using only Comet results (see Venn diagram). How can this happen? The cut off probabilities corresponding to protein FDR <1% for Comet (0.897393) and X!Tandem (0.872656) were as expected lower than combined (0.966806). 

Can anyone explain me why the combined library contains less proteins?

Also, I would like to ask you how to remove shared peptides from the library without getting format issues in further workflows?

Thank you in advance. 

Regards,

Nikita

[Eric Deutsch]

Hi Nikita, perhaps someone has some better answers than I do, but here is what I can say based on your description.

 

Regarding the cutoff, I wonder if in your combined analysis you are using a peptide-level FDR of 1%, whereas in the single analyses you are using an PSM-level FDR of 1%? A peptide-level FDR of 1% is more stringent than a PSM-level FDR of 1% and this might account for the difference?

 

I am not aware of a mode to remove shared peptides from a library, although this does sound like a useful feature. Maybe someone else knows?

 

Regards,

Eric

--
You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spctools-discu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/spctools-discuss/2da1064d-e4bb-41ea-9fa1-553c3d27c580%40googlegroups.com.

[Nikita Boeren]

Hi Eric,

Thank you for your quick reply.

I am using Mayu with 1% FDR on PSM level (mFDR) for all three libraries. I also filter the list to get the protein level FDR of <1% with -protFDR=0.01:t. The values I mentioned are extracted from this list (lowest score). 

As far as I know, with these settings I am using a protein level FDR estimation of <1% (on all three). Is this the correct way I am using it?

The search parameters for both engines are the same and other parameters are also kept the same for the libraries.

I think a tool to remove shared peptides would be very useful! Since I did not find anything about this how to remove those, I am wondering how people 'normally' do this. Some papers mention a total of x proteotypic peptides and proteins, thus my guess is someone must know how to do this. 

Regards,

Nikita

To unsubscribe from this group and stop receiving emails from it, send an email to spctools...@googlegroups.com.