Description
Assuming that the arrangment of the data is the same in your figures, so that rows are metabolites and columns are samples, then looking at the script for the largest_complete_subset, it looks like this function causes misalignment between the samples and metabolites. For example, if metabolite A once belonged to sample 2, and metabolite A is removed to the end of the row because its value was NA, then instead of sample 2 having its own metabolite A (which is an NA), it would have sample 3's metabolite A (not NA) and then the last sample would have sample 2's metabolite A (NA). Am I correct in my thought that all of the values in that row would shift to maintain the matrix structure, causing misalignment between the metabolites and the samples that they came from? Or is the data transposed compared to the figures from this paper?