Lines in the rectangle sometimes render as two-pixel wide instead of one:

If a single spectrum survives filtering and the the filter range is expanded so that more spectra become available, the navigation buttons stay dead:

In spectra that are in profile mode (i.e. not centroided), the n top peaks to be labelled are mostly points from the same peak. The n maxima to be chosen should be at least +/- 0.5 Da apart.

XIC filtering should rather be precursor m/z +/- 0.5 Da, with the latter settable as an option.

Selection region coordinates are slightly off when zooming in chart area.

At the moment the options are "mzR supported files", "RData files" and "all files". The latter two do nothing worthwhile. I suggest disabling "all files" and opening the object browser with objectsd within the file if the user selects an RData file.

Should be either removed as redundant or reconfigured.

It could be set as an option to display retention time in seconds or minutes.

Wrap existing code in msGUI methods.

Use type="h" when plotting the chromatogram and XIC.

There are many no visible binding for global variable notes when checking the package. Looking are several, they do seem do be false positives (like all the accessor functions defined in env), but the notes should be investigated in details and cleaned out if possible.

It would be useful to be able to display multiple XIC (using different colours). One would probably need a setting in the option window, with default to 1, where one could set more XIC filtering to be plotted. The extra precursor m/z +/- 05.Da should probably be set in a different window to avoid cluttering the main GUI.

Just tried to install msGUI on a clean machine and running msGUI complained about cairoDevice not being available.

I think toggle buttons look good for MS level selection. Also, when a user tries to unclick all levels, one level is automatically selected. This seems more natural with toggle buttons than with checkboxes.