Improve read groups with preprocessing markers

[sroener]

Modify existing read_group marker by prependen M_ for merged reads and T_ for trimmed reads. Everything else should not have a prefix

[makirc]

If you try to implement that through read groups, you need to make sure that all reads still belong to the same library & sample in the BAM header. It seems more effort. We just tagged it on the read IDs as this information is not really evaluated anyways (first part is typically just the name of the instrument and run date, but this is not consistent across sequencing centers). It could also be implemented with BAM read tags. Overall, I don't think this should have any priority.