Read inputs from tables

The prepare and map workflows currently use a dictionary approach to specificy genome resource- and sample-specific parameters, respectively. This is very tedious.

Refactor inputs to the Snakemake pipeline such that tables are read as inputs for different genome resources and samples.