Hi Chris,

The ImportPeakBoundaries feature was not designed to import MaxQuant data (mainly because we haven't received direct requests for this), but we'd very much like it to be able to do so, and in it's current form it may already work or be very close. Could you send me an example of the data you are trying to import (damodei at gmail)?

In answer to your specific question, the four columns you mention are necessary, and PrecursorCharge is optional. If you don't specify precursor charge, then the boundaries will be adjusted for every precursor in the skyline document under the peptide you've specified (e.g. if you have a line that looks like (VLVLTR, file_name_example, 40.0, 40.6), and your skyline document has a 2+ and 3+ precursor under VLVLTR, then skyline will adjust both of the precursors to (40.0, 40.6) boundaries. If you add the precursor column, then you can adjust the precursors separately, or only adjust some of them.

I actually have written a Skyline tip explaining all the details of how to do peak boundary import, but am waiting to release it until the next Skyline-daily is released. I'm hoping to add MaxQuant import to its capabilities -- if you send over your document, I can help you to get it working and make it as easy as possible.

Dario

Hi Chris,

I look a look at your Skyline document and evidence.txt file. The format is relatively close to what Skyline can read, but needs some tweaking. For now, you will have to tweak the file, but I hope to make Skyline able to just read these files without any change in the very near future. It looks like I had to change the following things:

1. The following column names needed to be renamed:

Modified Sequence => PeptideModifiedSequence
Calibrated retention time start => MinStartTime
Calibrated retention time end => MaxEndTime
Raw File => FileName

2. The form in which modified peptides were stored, e.g. _(ac)AAAAAAGPEM(ox)VR_, cannot be read by skyline. I changed the format to A(UniMod:1)AAAAAGPEM(UniMod:35)VR. That is, I did a find/replace to remove all the underscores, to change (ac) to (UniMod:1) and (ox) to (UniMod:35). Other acceptable formats are M[+16.0] for methionine oxidation.

3. The file as given included a lot of peptides that are not in the skyline document. To import peak boundaries, every line in the file must refer to a peptide and file in the skyline document. Even when I removed the lines from files other than e140117_HI820AS_01, there were still a bunch of peptides in the evidence.txt that were not in the document. Do you have an idea how those peptides got into the evidence.txt, or why they are missing from the Skyline document?

4. As mentioned above, only Skyline's "best" peak is changed on import. That is, if several lines contain the same peptide/file with different peak boundaries, then Skyline will just change the best boundaries over and over; the boundary that gets chosen will just be the last one in the file. Thus, you have to do something to remove boundaries other than the best from the evidence.txt.

All of these issues can be fixed in the .txt file, though with a moderate amount of hassle. Want to try this and see if it works? It seemed to work for me.

Also, some of the issues 1-4 can be easily fixed on the Skyline side. I just fixed #1 as I was writing this email -- it will show up in the next Skyline daily. #2 could be fixed if I understood more about MaxQuant's modification notation -- I've never used MaxQuant in my research so I'll ask Brendan or Mike about this. #3 isn't a problem with Skyline, in my opinion -- I think it's the user's responsibility to not have superfluous peptides in their text file. #4 could be fixed with some effort -- I'd have to add a column to specify if a peak is the "best" or not, and then Skyline could adjust both the best and non-best peaks. This is a bit involved, however.

Thanks for giving us the opportunity to test out MaxQuant imports on Skyline! Excited to make this work for you and other MaxQuant users.

Dear all,

I also tried to import MaxQuant peak borders. With the given workaround this worked well, thanks for that!

However, for using it as a routine it is not so convenient ;)

A more severe aspect is the issue 4, damodei mentioned.

This is, that MQ sometimes reports more than one peak boundary for one peptide with the same charge state.
For the MQ based quantification this is not so much an issue since "best peak" usually contains more than 90% of the total intensity and MQ takes the sum of all peak areas. But for skyline based quantification it is an issue. As damodei mentioned, skyline only choose one peak boundary and since the last peak is commonly not the most intense one e.g. elution in the high ACN range of the chromatography, this leads to a faulty quantification and to a wrong determined RT.

In my view MQ is quite good in identification and ms1 peak detection; skyline is a great tool for later data processing steps and for viewing the data. It would be great if it would be possible to couple MQ with skyline in a more proper way.

Are there any plans to do so?

Best regards
Fabian