Alignment of replicates originating from different ionization modes

Would it make sense to use the MapAlignerPoseClustering and FeatureLinkerUnlabeledQT nodes on files originating from both negative and positive ionization mode? Or is it safer to use these them first on negative files and then separately on positive files?

Generally, I would expect there would be completely different retention times for the same metabolite in negative and in positive mode. So, maybe it is better to align the negative and then separately the positive?