Update mass_terms.rst

[henryhlam]

Trying to clarify the usage of the mass terms in mzSpecLib. Some observations:

• We do not have terms for "experimental" precursor masses (but we have those for m/z). Obviously the MS measures m/z, not mass directly, so it may not be necessary. But some tools may calculate an "experimental" precursor mass from the experimental precursor m/z together with a determined charge state. And some tools may want to report a precursor mass deviation (i.e. experimental mass - theoretical mass) in the Interpretation section, which we also don't have a term for. (We only have m/z deviation terms.)

• I noticed that the QC folks have defined a few terms like: "MS_4000178|precursor ppm deviation mean" whose definition is: "the mean of the distribution of observed precursor mass accuracies ([observed mass accuracy) [in ppm] of identified MS2 spectra after user-defined acceptance criteria (FDR) are applied." Here, the "mass" deviation (which I guess is what it meant by "mass accuracy" distribution) probably should have been "m/z" deviation. I guess we won't use it in mzSpecLib but it may be a source of confusion.

• We also have not been clear about isotopes. The "theoretical neutral mass" can still be ambiguous because it can be (i) the mass of the monoisotope (which is what we typically calculate) of the analyte, (ii) the average mass (calculated based on natural isotope abundance) of the analyte, or (iii) the mass of one particular isotope that is being fragmented to yield the spectrum. For now, I stated that we meant (i) the monoisotopic mass, but it was not in the term definition.

• On (ii), the average mass of an analyte is often referred to the "molecular mass" (MW) by chemists. We don't have the term "average mass" and the term definition of "molecular mass" also doesn't say that it means the average mass.

• On (iii), there may be a use case that the MS selects one particular isotope to fragment. I don't know if I have seen a library spectrum like that yet.

[henryhlam]

After the discussion last week, I had some second thoughts, mostly surrounding the ambiguity about monoisotopic vs average vs isotopolog mass. It seems like we need better definitions.

Since we rarely targets one specific isotopolog to fragment anyway, let's ignore the isotopolog problem for now. Besides, in that case the "monoisotopic", "average" (though semantically those terms don't make sense any more) and the isotopolog mass would be exactly the same.

I suggest that we consider the following options:

1. Don't touch it and just use the existing terms. We just clarify the two most useful ones: "adduct ion mass" and "theoretical neutral mass" as monoisotopic masses when used in mzSpecLib, even though the CV definition does not prescribe it. This is what we want for now, and the change I made in the PR reflects this. In other words, we leave it open so that if in other formats they want to treat these terms differently, they can do so.

2. Redefine the terms "adduct ion mass" and "theoretical neutral mass" to be more specific what we want them to mean, i.e. monoisotopic mass. I originally thought this may be better, since these are in an MS CV (not some generic CV), and the terms are so fundamental to MS that we should have defined them more precisely. But if these terms are used in other contexts, they may be used to stand for different meanings. So "hijacking" them to mean what we want may cause some problems.

3. Define some new terms to use in mzSpecLib, so that we don't need to touch the existing terms which may be used in other contexts. For example, we may want to define:

adduct ion monoisotopic mass
adduct ion average mass

theoretical neutral monoisotopic mass
theoretical neutral average mass