Selecting SRM Targets Part 3

MSstats works!

After a few weeks, I finally figured out how to use MSstats to identify differentially expressed proteins and peptides. This Jupyter notebook has my workflow, but here are the highlights:

Step 1: Export data from Skyline

At first, I was exporting data from Skyline to meet the specifications for dataProcess in the MSstats manual. By following that route, I ran into an error saying that I had duplicate proteins and needed to specify a “Fraction” value. Better yet, there was no information about peptide fractions anywhere relating to MSstats! After a bunch of Googling, I found this handy function SkylinetoMSstatsFormat. I exported three separate Skyline reports to use this function, but with different BioReplicate and Condition information: 1) Bare vs. Eelgrass, 2) Sites only and 3) Sites and Eelgrass condition.


Figure 1. Columns needed for SkylinetoMSstatsFormat.


Figure 2. I saved the columns needed for SkylinetoMSstatsFormat under Export » Report » “SkylinetoMSstats”.


Figure 3. Example of Condition and BioReplicate settings, this one for both Site and Eelgrass condition pairwise comparisons.

Step 2: Use MSstats

The commands I used to conduct pairwise comparisons: SkylinetoMSstatsFormat » dataProcess » groupComparison

I then used groupComparisonPlots to graphically represent my data.

I found no differentially expressed proteins when doing pairwise comparisons between Bare and Eelgrass sites, but I did find differentially expressed proteins when looking at just sites, and when accounting for both site and eelgrass conditions!

Next step: examine annotations to see what kinds of proteins are differentially expressed.

Written on July 4, 2017