Request features: flagging (or solving?) saturated features

[lfnothias]

Hi,

Saturation of the detector in mass spectrometry is quite common in metabolomic. Obviously, these saturated MS1 or MS2 features cannot be compared properly between samples. Decreasing the sample concentration allows to avoid saturation of most abundant features, but will lead into a loss of sensitivity regarding the smaller intensity features, that deserves to be investigated as much as the most abundant ones.

So it would be great to tackle this matter, and I can see two options:

• The easiest solution, would be to indicated saturated ions in the output table of Optimus. Saturated peaks can be spotted easily by looking at their top-flattened shape of the monoisotopic ion, but also by looking at aberrant change in the isotopologue ratio.
• An idea to overcome this saturation issue, could be to use the 13C isotopologue (m/z +1) as a proxy to estimate the intensity of saturated monoisotopic ion across various samples. I know that Optimus contains an OpenMS option that pool detected isotopologues for a MS1 feature. But I think it takes into account only the m/z +1 (and/or +2), only if, the isotopologue ratio match a "natural" profil for metabolite. So that doesnt solve the problem here.

I take the chance to express how fantastic Optimus workflow is ! Thanks guys for sharing it with the community.

Louis-Félix

[iprotsyuk]

Hi @lfnothias, thanks for the report! From your experience, what's the minimum RT duration of such a top-flat part of the chromatography when you can say that the corresponding peak is oversaturated?

[lfnothias]

Hi @iprotsyuk

This saturation level will depend on the instrument and the methods. Here is an example of strong chromatographic/mass spec saturation. Column is phenomenex 2.1 mm x 100 mm (1.7 um). The instrument is a QTOF Maxis Impact. With the method I am using, the saturation level of this MS is around 1E7 . See the feature at mz 822.3433 (BPC of m/z 822.3 in violet, and BPC of all MS in red). Screen Shot 2016-10-18 at 2.35.55 PM.pdf Screen Shot 2016-10-18 at 2.35.44 PM.pdf

Regarding MS saturation in detail, here is the corresponding ions. Screen Shot 2016-10-18 at 2.31.05 PM.pdf Screen Shot 2016-10-18 at 2.31.18 PM.pdf Screen Shot 2016-10-18 at 2.31.37 PM.pdf

And here is a typical non-saturated feature we see. Screen Shot 2016-10-18 at 2.31.48 PM.pdf

Thanks and let me know if I can give more info.

[iprotsyuk]

@lfnothias, it looks like the identification of oversaturated peaks in m/z space will require the processing of profile-mode data which isn't acceptable by Optimus at the moment. From this of view, it's easier to implement the spotting of such features based on their distorted chromatographic profile. In the example you shared, the chromatographic plato lasts for about 10-20 seconds. Is it usual for oversaturated peaks or they can be shorter, down to several seconds?

[lfnothias]

Hi @iprotsyuk, Even in centroid mode, it should be possible to assess saturation by monitoring the variation in the isotopic ratio between the beginning of the chromatographic peak and the apex. The spotting of the saturated features based on chromatographic peak shape could do the job so we can parse these features for statistical analysis of the data. However, I am concern about how it could deal with the variety of chromatographic conditions (columns, flow, particules, ...).

LF