autoQC explicit retention time window

autoQC explicit retention time window halwaseem  2024-02-06 10:26

I am running serum samples using an MRM method on an Agilent triple quad. Only product ion data is collected (at specified RT windows). There is no full MS data. I would like to use autoQC to monitor the signals of the deuterated internal standards spiked into a quality control sample. I have tried including and excluding the explicit retention times but Skyline isn't correctly integrating the ISTD peaks due to inter-batch retention time shifts(~0.2-0.8 min). Even with a window that captures the shift, Skyline is going in an peak picking noise at the explicit retention time instead of the nicer peaks.

I have attached 3 examples of 3 different compounds in the autoQC transition list that are being properly/improperly integrated.

Since this is for SST, I would prefer for this to be automated and not have to go in and manually adjust the peaks so that they show up correctly in Panorama. Is there something in the settings that can be adjusted to address this issue? Is there a way to set a minimum peak intensity for autoQC integration? There are many examples where there is an obvious peak in the RT window, but Skyline integrates noise instead.

If a iRT calculator was added to molecule settings, would Skyline be able to apply that as raw files came in via autoQC?

Nick Shulman responded:  2024-02-06 11:34
When Skyline detects a peak which overlaps with the Explicit Retention Time, Skyline will choose that peak no matter how low quality that peak is. The only way to avoid this behavior would be to leave the Explicit Retention Time blank. The Explicit Retention Time Window is only used if no peak was found that overlaps with the Explicit Retention Time.

Fortunately, with your dataset there might be other things that you could do to give Skyline different hints about which is the correct peak.
I see that you are extracting chromatograms for two transitions. If you had a spectral library, then Skyline could use the spectral library to know what the expected ratio of those two transitions' peak areas should be.

If you have a Skyline document where you have chosen the correct peaks you can use the menu item:
File > Export > Spectral Library
to create a spectral library with pseudo spectra in it where the intensities are based on the integrated peak areas in your Skyline document.

You can then use that spectral library in another document, and Skyline will be able to calculate a "Library Dot Product" and that will be part of the score that gets used to choose the best peak.
It would be a little tricky to get this to work today in Skyline 23.1 or the latest released Skyline-daily because those versions do not calculate the library dot product unless you have at least three transitions, so you would have to add a third transition whose m/z was .0001 units different from a transition that you already have.
Things will be a little simpler when the next Skyline-daily comes out in a few weeks because we recently removed that limitation and that future Skyline-daily will calculate library dot products even when there are only two transitions.

If creating a spectral library does not improve things for you then I would recommend using iRT where you tell Skyline about some molecules that are very easy to reliably detect and their location is used to determine where to find the rest of the molecules.

By the way, if you want to see why Skyline has chosen a particular peak you might find some useful information in the "Candidate Peaks" window which you can find at:
View > Other Grids > Candidate Peaks
-- Nick
halwaseem responded:  2024-02-07 06:36
Thanks Nick for all this info.

Candidate peaks was very interesting information and that led me to Refine--> Reintegrate, and then I edited the default model to only use intensity and peak shape. Skyline did a better job choosing the correct peaks.

I don't quite understand the candidate peak scoring. I had many cases (pre- reintegration) where the chosen peak scored really high in terms of weighted intensity and peak shape, but it doesn't even look like a peak. I have attached an example an included a screenshot of EIC.
Nick Shulman responded:  2024-02-07 07:28
I can't tell from the screenshot which molecule and replicate you were looking at.
Can you send me a screenshot of the whole Skyline window so I can see exactly what is selected in the Targets tree and the Replicates dropdown at the top of the Targets tree?
Also, you should probably send me your Skyline document again (File > Share to create a new .zip file) just in case your document is different from what you sent me before.
-- Nick