Replace current organelle search by full assembly algorithm

[dirkjanvw]

I just discovered that with oatk/mitohifi it should be possible to assemble mitochondria (and chloroplasts) instead of only identifying them in an assembly. I will do some testing to see how to connect this into the currently assembly model while retaining the modularity of assembler choice. There are two options to investigate:

1. do oatk/mitohifi assembly in parallel to full genome assembly and then after that filter contigs
2. do oatk/mitohifi assembly first, then filter reads, then do nuclear genome assembly

Some practical notes to future self:

• oatk can be installed through bioconda but mitohifi needs to be run via apptainer.
• I did a very quick test with some A. thaliana data and it seems oatk on the input hifi reads is not significantly faster to run that oatk on a resulting hifiasm assembly; meaning that probably we only need to investigate the former.