Electron mass shift for calculated masses

[wkumler]

As Dong reported two years ago over at IngallsLabUW/MARS_Project#1 and I've confirmed today, using the mass calculated by Rdisop directly results in a pretty persistent mass shift the size of an electron. Regina implements this correctly with the getExactMass script in 8fb6679 by subtracting the mass of an electron when in positive mode and adding an electron back when in negative mode.

This is mostly caused by the "Empirical_Formula" including the adduct but NOT the charge. A simple test of Rdisop shows that the PubChem formulae are consistently adducted by the mass of a hydrogen ATOM, not a hydrogen ION:

And this difference is especially striking when considering how large the gap is to the "proper" mass:

Feels like something we should fix but Rdisop doesn't handle charged ions at all

[wkumler]

Here's the comparison to Skyline's calculated masses for positive mode:

and negative mode:

for the Falkor data

[wkumler]

I ran through and calculated all the masses using the new formula and ended up with some ambiguous feedback, as we suspected form the histogram above. The PDF I've attached shows the new mass vs the old mass plus some ppm windows for context and it doesn't seem like this is a unilaterally good idea. Not sure where the error's coming from but it makes me want to push off this change for later. at least.
new_mass_checksheet.pdf