Calculate reference proteome similarity when running `generate_protein_fasta` commands

#450 adds calculation of reference proteome similarity metrics to the main pVACtools pipelines. We should also add them to the individual generate_protein_fasta commands.

This needs to be made configurable because the underlying modules are also used by the generate_combined_fasta steps in the main lib/pipeline.py module. We don't want the proteome similarity metrics to be calculate here because they aren't needed and will add a lot of overhead. This could be achieved by adding a --no-metrics flag to to the commands, which can then be set when calling the modules in generate_combined_fasta. This would also allow us to make the calculation of manufacturability metrics optional for these commands. Manufacturability metrics are currently always run and, thus, get extraneously generate in generate_combined_fasta where we immediately deleted them.

Start by tracing the individual generate_protein_fasta commands and the generate_combined_fasta steps in lib/pipeline.py, then inspect how the underlying modules calculate proteome similarity and manufacturability metrics. The work is done when both metrics can be controlled for the individual commands without adding unnecessary calculations or generated files during the combined FASTA pipeline.