• Sign In
  1. Requests

targeted metabolomics with Skyline

support

View Request

view list print
targeted metabolomics with Skyline arnaud garcia  2018-02-08 10:44
 
Dear Skyline developpers,

I work on a project of using this software to extract feature from Orbitrap raw data to obtain a standardized profile of metabolites. The goal is
- Upload raw data files
- Upload a database of molecules to screen for ( with exact masses/ formula and known retention time)
- Extract from the database the area of the masses that are found only in the time windows (and if there is nothing, then the area should be left blank).
- Get in an excel file (in the best scenario) the observed area, peak info, masses and retention time for each molecules screened.

After uploading the list (using the "Edit > insert > transition list") and the insertion of raw data, it occurs to me that I couldn't get the selective windows of retention time correctly even after "re-scoring" of the data and changing parameters. For many metabolites it was good because they were found within the right range. However (fig 1 in the file attached), many ions were returned by the software to be matching our molecules of interest while they were out of the suggested ranges.
Same, another problem was the mass accuracy of my list. I was working with an Orbitrap format and couldn't reach manually the mass accuracy settings ( from Settings > Transition Settngs > Full Scan). The accuracy at my selected settings were too high for my study so i had to switch my precursor mass analyzer to "centroided" in order to manipulate the mass accuracy, even if it is less than ideal.

So here i have two main questions :
- is it possible to integrate areas automatically for a set m/z windows and a set Retention time windows regardless of the kind of data used (orbitrap, Waters, Agilent..)?
- Is there a way to select easily the mass accuracy of the selection of masses when selecting MS1 Orbitrap files?

Retention time accuracy and mass accuracy are really important for selectivity in metabolomics and since the goal of Skyline is also to cover small molecules integration, those parameters are still confusing to me.

Thank you very much and have a lovely day,

Arnaud
 
 
Nick Shulman responded:  2018-02-08 15:56
If you want to force Skyline to choose a peak within the explicit time window that you have specified, the way to do that is to tell Skyline to only extract the chromatogram for that time range.
That is, you go to:
Settings > Transition Settings > Full Scan
and under "Retention Time Filtering" choose "Use only scans within 0 minutes of predicted RT".

I believe we currently always recommend that you use "centroided" for your "mass analyzer". The reason for this is that the vendor's algorithms do really smart things with the shape of the observed peak. If you do not choose "centroided", then Skyline sums the profile intensities across a m/z channel whose width is determined by the resolution settings that you specified.

When Skyline extracts chromatograms, it does record the mass accuracy (i.e. the difference between the peak of the m/z channel and the expected m/z). Unfortunately, this mass accuracy value does not get used by Skyline when Skyline is deciding which is the best peak. In order to get Skyline to use the mass accuracy value in peak picking, you have to use an mProphet model.
I believe there is no way to train an mProphet model in a small molecule Skyline document (partly because we have not figured out what a small molecule decoy is), but there might still be a way to tell Skyline to use a scoring model that takes into account mass accuracy.

It would be helpful if you could send us your data so we can look at it.
In Skyline, you can use the menu item:
File > Share > (complete)
to create a .zip file containing your Skyline document and supporting files.

You can upload that .zip file and some of your .raw files here:
https://skyline.ms/files.url
 
arnaud garcia responded:  2018-02-09 00:17
Dear Nick,

Thank you for the fast and furnished answer. I will try those asap. When switching the retention time filtering to 0 minutes of predicted Rt, does the windows of Rt is then limited to find the apex and integrate what is in the windows only, or does the window of integration gets a big wider to include the whole peak? In case of a peak near the border of the window, it is possible that the peak shape exceed the borders.

I will make the deposit later for the files for you to look at.

Thank you again,

Arnaud
 
arnaud garcia responded:  2018-02-28 06:41
Hi Nick,

I come back to you as I found an odd feature while trying to apply your advices. The mass accuracy problem was cured, using the "Centroided" option for "Precursor Mass Analyzer" and 1ppm that you proposed; it resolved the problem.
However, I can't get the Retention Time prediction right.
- Either I can import the predicted retention time and explicit window ("capture5") and the raw files with the "Retention Time Filtering" setting on "Use scan within minutes of MS/MS IDs", but the peak picking is not good as it integrate a peak which is out of the range, while there is a low intensity ion in the right frame that could correspond. This one is an example but I got a lot more of them.
-Either I can't import the raw file when I set the "Retention Time filtering setting" on "Use scan within minutes 1 of predicted Rt" as you advised (1 or 0 minutes of predicted Rt still give the same result). It then displays the error "Transition Setting Full Scan is set to use predicted Rt but no algorithm has been specified" (see "capture2"). As your advice was to use this force Skyline to choose a peak within the explicit Time window, I suppose there is something I miss in the way

I've joined the data with the Share option for you to look at it.

Have a nice day,

Arnaud
 
Nick Shulman responded:  2018-02-28 07:50
I believe I was wrong about being able to use "Explicit Retention Time" to limit the length of chromatograms that Skyline extracts.
I do not know what I was looking at in the code that made me think this would work. (Oops! Sorry about the incorrect information.)

We are in the process of getting retention time prediction to work for small molecules in this next version of Skyline.
Can you send me your .raw files? You can upload them here:
https://skyline.ms/files.url

One thing that looks like it would almost work is creating a spectral library from your Skyline document. That is, you could create a Skyline document with one replicate, and then perfectly adjust all of your peak boundaries, and then use the menu item:
File > Export > Spectral Library
and then create a .blib file.

That technique works great for telling Skyline about the expected transition intensity ratios for small molecules, but one thing that it will not work for is telling Skyline about predicted retention time. I believe that's just an oversight on our part, and we will be able to fix it soon in Skyline-Daily.

It would be helpful to take a look at your .raw files so I can make sure that this feature will work.