Add BowTie2 support

[jfy133]

Is your feature request related to a problem? Please describe.
This was previously in EAGER1, but also has benefit of "better for investigating absent gene regions since it can split the reads when mapping." (requested by Ash V. in CPH)

[jfy133]

Recommended maping settings for aDNA: https://academic.oup.com/mbe/article/35/5/1120/4844088

Additional for UDG treatment and looking for methylation: https://www.frontiersin.org/articles/10.3389/fevo.2020.00105/full#B4

[jfy133]

Ok, changed my mind on this one. We should include it in 2.2 because it would be important for eager1 parity and also given the Poullet and Orlando paper, would make it more relevent.

I will try to contact the authors and get their recommendations.

[apeltzer]

Docker building here in the repository could also be added then, if we anyways play around with the dev container adding bowtie2 now too

[apeltzer]

Bowtie2 dependency is added too. Remaining would be to allow for alternative mapper instead of bwa ;-)

[apeltzer]

This study reveals that the Bowtie2 mapping conditions recommended (sensitive option, end-to-end mode) returns with larger numbers of read alignments that increase the number of genomic windows available for inference as well as their sequence coverage, which improves accuracy of the predictions. This has important consequences for the nascent field of ancient epigenomics, which, to the best of our knowledge, based all previous predictions on BWA DNA alignments.

https://www.frontiersin.org/articles/10.3389/fevo.2020.00105/full

So we should enable at least that mode if possible :-)