Benchmark Mmseqs2 clustering based decontamination method with contscout fungi decontamination results.
As MMSeqs2 automatically picks the optimal clustering strategy based on the coverage mode it is smarter to play with the cov-mode parameter rather than the cluster mode.
- Download Fungal genomes of contscout
- Cluster repdb+these genomes (checking if some species are duplicated)
- Use different mmseqs params and see overlap with contscout results in supp table 4
snakemake -j 2 --ri -p -k --executor slurm --workflow-profile workflow/profiles/default/
You can look at the notebook in workflow/notebooks/repdb_decon.Rmd to see what we've found out.
- Why only fungal though
- decide what sequences to cluster in RepDB
- It could be interesting to see if intraeuka contaminants have something weird