local FDR/PEP

support
local FDR/PEP Tobi  2019-08-19
 

Dear Skyline team,

could you imagine including a local FDR value to complement the Q-values in the results grid? (PEP) For me it seems nice beeing able to retrieve the actual chance of a specific peptide beeing a false discovery, but you might know more about limitations and relevance of this.

https://pubs.acs.org/doi/abs/10.1021/acs.jproteome.7b00170

Looking forward to hear your opinion on this.

Best regards,
tobi

 
 
Brendan MacLean responded:  2019-08-29

Hi Tobi,
Really informative post. Thanks!

I have posted an issue to keep this on our TODO list, but I can't say exactly when it might get implemented. Even in just the area of peak scoring there are a few ideas I would prioritize implementing before this, like peptide- and protein-level q values.

Thanks for your feedback.

--Brendan

 
Brendan MacLean responded:  2019-08-29
 
Tobi responded:  2019-08-30

Dear Brendan,

thank you very much, sounds really good. I like the idea of protein-level q-values. If you get to this point, could you please add the option during peptide and decoy filtering against a background to treat Leu/Ile as the same (undistinguishable)?

Best regards,
tobias