loading the STAR mapping index takes long. Treat this as a service

[marvin-jens]

snakemake support 'services' https://snakemake.readthedocs.io/en/stable/snakefiles/rules.html#service-rules-jobs

which can be started and stopped. Sounds like a perfect fit for loading the STAR genome index and releasing it again after the last instance has quit. In the meantime, STAR mapping jobs do not have to load their own copy of the genome index which would make them faster to start mappting and reduce overall memory footprint