Right now the LARRY barcode matrix generated by the pipeline assigns a random number to each barcode. This becomes a problem if you are running samples from the same LARRY experiment in different pipeline instances. The same barcode sequence will probably have different IDs across both samples, which is a pain (and not very smart).