Add new de novo prediction features for wastewater and mixed samples

[hoelzer]

This is a completely wip PR while working on integrating additional features as partly documented in #275 .

So far, the idea is to

• evaluate/finalize frejya as reference-based approach
• adding additional tool(s) for de novo detection of amplicon clusters and associated mutation profiles (e.g. via devider and the covariant function of freyja)
• and finally comparing the de novo detected mutation profiles w/ known ones from patient samples to ring a bell if something potentially new/more diverse pops up