Add genome/reference definition section to configuration

[tedil]

Currently, steps/tools define their own reference naming conventions.
This can lead to incompatibilities between steps, accidentally selecting incorrect references etc.
Unifying the way references are handled could help prevent that.

However, tools sometimes rely on specific aliases. Therefore, we could add support for aliases, something along the lines of:

reference:
  name: GRCh38
  aliases:
    - ensembl: GRCh38
    - ucsc: hg38
    - panel_of_normals.purecn: hg38

In principle, we could also directly specify the source and release directly, such that the reference sequence can be downloaded (and cached) automatically, e.g. for ensembl:

  reference:
    […]
    source:
      ensembl:
        species: homo_sapiens
        release: 111
        build: GRCh38

or local:

  reference:
    […]
    source:
      local:
        path: /path/to/ref.fasta
        species: homo_sapiens
        release: 111
        build: GRCh38