Exporting all peaks and mass accuracy for chromatograms thomas hopf  2017-07-24
 
Dear Skyline developers,

I am trying to set up an automated evaluation scheme for peak picking outside of Skyline. For this, it would be very nice to have access to all candidate peaks and their associated mass errors. While I see that information in the GUI, I wasn't able to fully access the underlying data so far.

1) When exporting the chromatograms for all of the precursors, besides the intensities is it somehow possible to get the corresponding mass error/accuracy for each time point? In the reports, I only seem to be able to get the error for the one chosen peak per precursor.

2) For each precursor, is it possible to get a list of all candidate peaks chosen by the Skyline heuristics?

3) I seem to be able to extract some of the above information by exporting the mProphet peak features; however the associated mass errors for the mProphet candidate peaks seem to differ quite substantially from nearby Skyline peaks. How does this occur?

4) Can the mProphet features be automatically extracted using a command-line parameter?

Thanks for your help!
 
 
Brendan MacLean responded:  2017-07-29
Hi Thomas,
It doesn't sound like we quite have everything you are seeking. You should definitely have a look at both:

File > Export > Chromatograms
File > Export > mProphet Features

The latter should get you all feature scores for all available peaks, assuming you don't check the "Best scoring peaks only" check box. The former gets you all the raw chromatograms.

It wouldn't be hard to add an option to the chromatograms export to export our mass accuracy calculation for every point in each chromatogram. That is how we store it to be able to calculate mass accuracy for any peak a user might integrate manually.

It would also not be difficult to make either of these accessible from the command-line. Though, neither of them are yet.

How confident a programmer are you? Skyline is, of course, open source, and we would welcome your effort to improve these features to more fully match your needs. And, if you work out a way to improve on the current peak picking, we would welcome your help in integrating the improvements into Skyline itself.

Note that we also have File > Import > Peak Boundaries and internally we can perform this operation orders of magnitude faster, if you are just picking from the list you would get from File > Export > mProphet Features. If this matches your intended use case, we might want to implement a way to specify peak indexes in File > Import > Peak Boundaries, instead of only start and end integration times, which is more general, but much slower for Skyline to achieve.

Thanks for your interest in working on chromatogram peak picking. Let us know how we might help.

--Brendan
 
thomas hopf responded:  2017-08-01
Hi Brendan,

Thanks a lot for your detailed reply!

I should be able to add the necessary functionality - for that I'd greatly appreciate if you could give me a rough pointer where the export and command line functionality lives in the source code, and what data structure to look out for that holds the mass accuracies.

I was also wondering if there is a way to limit the peak detection to certain precursors, e.g. the heavy standard? This would be very helpful for situations when we are measuring the standard on its own.

Thomas
 
dsullivan responded:  2019-09-12
Hi Brandan et al,

I am wondering if the Import > Peak Boundaries can be accessed from the command line yet?

Regards,
Devin