Fragment formula annotation is incorrect for multiply-charged ions

[meowcat]

For charges > 1 (e.g. mode = "m2H_c2"), fragment formula annotation is incorrect.
The m/z used as input for generate.formula is corrected by # electrons corresponding to the precursor charge, but not multiplied by z. This leads to masses that correspond to singly-charged fragments that are off by one electron (for the case of z=2).

In reality, the observed fragments would be either singly or doubly (or triply, etc) charged, and for each possibility both the electron mass and z need to be adjusted.

RMassBank/R/leMsMs.r

Lines 643 to 658 in 7b780e2

	peakmatrix <- lapply(split(shot,shot$row), function(shot.row) {
	# Circumvent bug in rcdk: correct the mass for the charge first, then calculate uncharged formulae
	# finally back-correct calculated masses for the charge
	mass <- shot.row[["mz"]]
	mass.calc <- mass + mode.charge * .emass
	tolerance <- ppm(mass.calc, ppmlimit, p=TRUE)
	if (mass.calc > maximal.mass + tolerance)
	return(t(c(row=shot.row[["row"]], intensity = shot.row[["intensity"]], mz=mass, formula=NA, mzCalc=NA)))
	peakformula <- tryCatch(
	suppressWarnings(generate.formula(mass.calc, tolerance,
	limits, charge=0)),
	error = function(e) list())
	# was a formula found? If not, return empty result
	# ppm(mass, ppmlimit, p=TRUE),
	# limits, charge=1),
	#error= function(e) list())

[meowcat]

Working on it in branch https://github.com/MassBank/RMassBank/tree/multiply_charged. Since this changes a core function, I am hesitant to push it into 3.15 release