Peak Scoring with heavy peptides

support
Peak Scoring with heavy peptides tilman werner  2021-01-08
 

Dear Skyline Team,

Happy new year - thank you for this amazing tool!
I have a question regarding peak scoring. I have run a PRM experiment with spiked in heavy peptides to detect several cytokines. Some we could clearly detect, but with others the results are shaky and we are not really sure if we should count the peptides as detected in our sample or not. The main issue is which peaks to already count as a fragment pattern and which ones to discard as noise.
I was wondering if there is a tool in Skyline that allowed us to calculate something like a Q-value for the peaks by using the heavy peptide peaks as input/template instead of a seperately generated peptide library? Or is there another way to discern signal from noise in the presence of heavy peptide standards?

Thank you for your great work: Skyline is by far the fastest way to go from measurement to shiny result!
Best wishes,
Tilman

 
 
Brendan MacLean responded:  2021-01-13

Hi Tilman,
Generally in the presence of heavy standards, Skyline scores the heavy standard and then just quantifies the signal on the light channel. Skyline has background subtraction, which should reduce the noise contribution considerably, but it can also be advisable to perform some assay development ahead of time to determine which transitions may have interference and therefore should be marked as not quantitative, which would keep them from contributing to the TotalArea or light:heavy ratio.

It is generally enough to prove you are integrating the right retention time range based on consistency of your detection of the across replicates, based on intensity and retention time.

If you really want a q value, you can look into using mProphet in Skyline, described in the Advanced Peak Picking tutorial, but this is now considered most appropriate for larger-scale methods like DIA, and usually without injected standards.

Hope this helps. Sorry for not getting back to you sooner.

--Brendan