Method needs major changes in upstream projects

scMerge has two different modi, running with a list of precomputed "stable" genes (available for human and mouse only) or computing a list of stable expressed genes from the input data. To run with the first mode the dataset organism needs to be included into meta data? How to filter non human/mouse datasets? The second mode is computationally expensive. Maybe this step could be implemented in a way that it is only run once per input dataset and not per parameter run? Overall this method needs major revisions and will be postponed for now. Please approve!