Skyline DIA output processing in MSstats

[Giovanni Canil]

Dear MSstats group,

I was wondering how MSstats processes data coming from Skyline DIA analysis. 

If I understood correctly, the Area of the fragments identified by Skyline for a specific peptide  are added together to obtain the peptide intensity, and all the peptides intensities for a specific protein are added  together to obtain the protein intensity, am I right? (I believe here the  'intensity' means the area without additional processing?)

Finally, what if I have several peptides coming from Skyline, some of which are well defined by many fragmentations but some are suffering from interferences? Is there a possibility that MSstats filters away these peptides or is it something that has to be done previously in Skyline, playing on the dotp value?

Thank you very much for your time and the development of such a useful tool,

Giovanni

[Devon Kohler]

Hi Giovanni,

We actually perform the analysis a little differently than how you describe above.

For DIA we take each individual fragment as feature which represents the underlying protein abundance (i.e., we are not grouping feature by peptides). All of these features are used individually and summarized using Tukey's Median Polish (similar to median summarization) in the dataProcess function.

In terms of feature selection (removing features with interferences/missing values/ect) there are a couple options. In the dataProcess function you can choose to use all fragments, the top N fragments (based on average intensity), or have MSstats select the "best" features using an internal selection algorithm. There is more info on these options in Box 2 of our new MSstats paper : https://www.nature.com/articles/s41596-024-01000-3.

Hope this helps!

Devon