search

HUPO-PSI / mzTab

mzTab Reporting MS-based Proteomics and Metabolomics Results
https://hupo-psi.github.io/mzTab
39 stars 17 forks source link

Question: 6.3.10. exp_mass_to_charge #100

Closed timosachsenberg closed 6 years ago

timosachsenberg commented 6 years ago

Is reporting the average (over assays) experimental mass for the primary adduct (see exp_mass_to_charge 6.3.10.) really useful information? If so, can we provide some additional information or an example (e.g. if multiple samples are measured in several assays)?

timosachsenberg commented 6 years ago

e.g., if and how this relates to retention_time_in_seconds? (in one we talk about a Master or aggregate run, while in the other we talk about averaging assays)

kayrein commented 6 years ago

I think the critical point here is that the mean of the experiment m/z ratios is meaningful in a way that the mean of retention times isn't.

Experiment m/z experiences random measurement variation, and the mean of the measurements should be better than any individual measurement.

Retention times may drift as a column is reused, and then be reset somewhere else once it's eventually replaced. So taking the mean of retention times would bias the results if a metabolite isn't detected in all of the runs (which it often isn't) and be non-meaningful.

Retention time is more useful within a given run. A master, or pooled sample can be used as a representative. Alternatively if you have a created an aggregate of aligned runs then that is the obvious choice.

andrewrobertjones commented 6 years ago

We agreed to remove these two mentioned datatypes from SML level, on the basis that SML level should not try to report experimental data that should be captured in Features. References to SMF level should capture exp_mass_to_charge and RT values.