Fold change correlation plot is too bespoke

[Asplund-Samuelsson]

The fold change correlation plot fc_correlation.png made by the quality control script qc.R is a bespoke graphic that is hardcoded to work with one specific version of the dataset. It would be convenient if this script could make a fold change correlation plot for any input data. This would require extensive rewriting of the plotting code.

Example: if there are fewer than 19 metabolites, at least one metabolite set mets will be empty, resulting in an error in the make_plot function.