classify_kraken2_batched
Classify many unaligned BAMs with kraken2 and, optionally, keep or remove the reads that fall
within given clades. Loading the kraken2 database is the expensive part of a run, so this
workflow loads it once per VM instead of once per sample.
How it works
Samples are split into batches of at most samples_per_batch. Each batch runs on one VM that
- extracts the database once,
- reads it into the page cache once,
- classifies
concurrent_samplessamples at a
This is a companion discussion topic for the original entry at github.com/alex-omics/microbial-genomics-wdl/classify_kraken2_batched