Proteomics pt. 3:

or why you should put informative error messages in your programs.

After getting the Pecan workflow working on Emu, we converted Roadrunner over from OS-X to Linux to have a second machine to run Yaamini’s data on. It seemed to be a straightforward thing, but trying to run a shortened set of her samples brought up an issue with error logging in Pecan.

When you run the entire isolation scheme via GridEngine it outputs a series of log files, none of which say anything went wrong exactly, you just don’t get any output files, which is frustrating. When you run a single isolation window instance it gives an error message PECAN.Chromatogram|INFO| 0 MS2 scans! Quit now which, at least to me, sounds like something is wrong with the data and there are no MS2 scans there. Apparently what it really means is there are no MS2 scans in the individual isolation window, but the data is really there. Not very intuitive.

After figuring that out, I started running 3 known good (as in it has peptides found) isolation windows on Laura’s data on Roadrunner, hoping to get enough output to run through Percolator and Pecan2blib which will allow us to have something for Skyline. Once we get that far, we can at a minimum test our workflow to make sure everything works.

Currently Pecan is Peaking peaks for the three isolation windows, so next up is Percolator. Hopefully a faster step than Pecan.