IMS Predictor questions Juan C. Rojas E.  2020-08-21

Dear Skyline team,

I have a few questions I hope you can help me clarifying.

  1. How does Skyline determine the CCS area from WATERS HD files? What equation is implemented?

I checked at some point if the CCS areas reported in Skyline matched the ones calculated with Driftscope and I noticed slight differences. Maybe due to small differences in the application of the calibration equation stored in the files?

  1. When the IMS predictor is run automatically with "Use Results" it will attempt to only use the areas inside of the RT boundaries?

I import my DIA data with small retention time windows to limit false integrations and then run the predictor. However, after running the predictor I noticed that the EICs show the full chromatogram and not only the small retention windows imported in beginning.

  1. Is there a bias for optimizing the IMS window of the fragments (i.e. high energy scan) above the one for the precursors (i.e. low energy scan)? Or is it some sort of weighed contribution?

I have tried using the automatic predictor and very often its determining incorrect offset estimations. In the image attached you can see that when automatic is ran it usually ends up with lots of positive offsets which is the complete opposite of what is expected from associated fragment ions in the HDMSE high energy scans. This happened because it chose same fragment ions that came from other co-eluting or closely eluting molecules.

Of course, with my manual corrections I still had some mistakes due to badly curated peaks, but at least the bias is towards negative offsets for most of them. I should mention that I use a low resolving power of 15 since is the value the encompasses all area for most of my molecules with my current instrument settings and analytes relative abundance.

Sorry if these were too many questions for one post, but I thought they were all related.


Brian Pratt responded:  2020-08-21

Hi JC,

Taking these in order:
1.Waters provides this calculation in their proprietary file reader library. It's exact implementation is unknown to us.
2. Skyline looks for the IM value which produces maximum signal at the chromatographic (Retention Time) peak.
3. It's the same approach for fragments - pick the IM value that gives maximum signal at the RT peak. It certainly sounds like it's being confused by interference from coeluting molecules.

It's important to understand that this technique necessarily assumes no interference. Ion mobility separation is an excellent way to deal with interference from coleluting molecules, but if it was possible to derive CCS from samples with interference - to ignore the interference without the help of IMS - then we wouldn't really need IMS in the first place. The way to develop a reliable CCS library from mass spec data is with simple mixtures such that there are not coeluting molecules of interest. After that you can throw everything into the LC at once and let the additional dimension of IMS do its work, but you do have to bootstrap it first.

As you probably know there's some interesting work being done in CCS prediction that may streamline this process, but we're not directly involved with that. You can import such values into Skyline, though.

I hope this helps, and of course feel free to continue the discussion here.

Thanks for using the Skyline support board!

Brian Pratt

Brian Pratt responded:  2020-08-21

All that being said, there's certainly a case to be made for not proposing high energy offsets that move in the wrong direction. I'll see about improving that.