Hi Hongbo,
I am certainly considering how to make the "Match tolerance m/z" value something more understandable to people, as I am seeing it commonly misinterpreted like this.
The "Match tolerance m/z" value is only applied to m/z values that are used in instrument settings (i.e. instrument methods), and never to values related to m/z values reported by a mass analyzer. For example, Match tolerance m/z is used to match the m/z values reported as the precursor and product m/z values in SRM and the precursor m/z reported in targeted-MS/MS. In all both cases, the instrument operator supplies a method with the m/z values in it, and the instrument simply isolates ions around the m/z values it is given, and reports them back in the resulting output.
If the method was created with a less than exact m/z value for a given peptide or product ion, then the resulting data will still report that m/z value, since that is what it measured in the cases listed above.
In several important data sets to the original design of Skyline, m/z values rounded to a single decimal place were used in the instrument methods (e.g. 572.6), and in some cases there were even errors in the rounding (e.g. 572.5441 rounded to 572.6).
This, then, is the entire purpose of Match tolerance m/z: to give Skyline a tolerance by which it matches human supplied m/z values with the far more accurate m/z values calculated inside Skyline.
The value has nothing to do with the accuracy of a mass analyzer in any instrument.
Hopefully that helps clarify why this value will never be changed to allow values in PPM, since I know of no instrument which allows users to specify target isolation m/z values with PPM accuracy.
Further, when Skyline extracts chromatograms from full-scan data, it tags the resulting chromatograms with m/z values at "float" precision in C#. That means 3 or 4-decimal places for typical proteomics m/z values. And that probably means that if you set Match tolerance m/z lower than 0.0005, Skyline may begin to fail to match certain chromatograms with the element for which they were extracted. I should probably look into this more closely, since I am currently allowing a minimum of 0.0001, which should still work for most things, but I believe I can imagine cases that should work that would fail with a match tolerance m/z below 0.0005.
Skyline itself exports transition lists and targeted m/z methods with m/z values to 6 decimal places, but regardless of what the instrument does with these numbers, I see now that Skyline will never carry this high precision through its entire pipeline.
I doubt this will effect many people, since even with your common misunderstanding, you are only trying to set the tolerance to 0.005 to 0.002, which should be fine for extracted chromatograms.
Thanks for bringing this up. I will certainly give more thought to how to make it clearer that Match tolerance m/z is not related to mass analyzer accuracy.
--Brendan |
Hi, Brendan,
Thanks for the clarification. And I want to ask if there is a certain mass accrary range Skyline uses to extract chromatorgam from raw files or there is a default value? Normally, when i did the EIC manually in XCalibur, I always set 5ppm.
Thanks
Hongbo |
Hi Hongbo,
Skyline does not use mass accuracy as an extraction parameter. It always extracts its chromatogram from a range equal to 2x the expected peak FWHM in the m/z dimension (i.e. the resolution). So, the resolution settings in the Transition Settings - Full-Scan tab control how chromatograms are extracted. No centroiding is performed, and currently mass accuracy is not even calculated, though it will be in the future. I do not foresee discarding intensity values in the future due to poor mass accuracy, but we will likely use mass accuracy as an important score in picking chromatogram peaks.
Hope that is enough information to be understandable. Very pressed for time today.
--Brendan |