Notes

To be updated

Requirements

• new(er/est) version of snakemake
• rgi: should now be able to handle * in FAA files
  • important: does removing * from FAA sequences affect some results or should that be done in any case (?)
• signalp, v6
  • paper
  • web site
  • repo
  • need to check compatibility
• optional: plasflow --> should alternatives be added?

Configuration

Config

• no runtime configuration (only in rules and profiles)
  • rm runtime, mem
• sample table (see also below)
  • ID, FNA, (FAA, FNA/FAA mapping)
• no project and datadir
  • either only one output path or use work folder via profiles
• paths to DBs: single path? group by attribute?
• allow multiple steps (see also below)

Sample table

• columns: ID, FNA, (FAA, FNA/FAA mapping)

Profiles

• different types:
  • generic (w/o a scheduler)
  • HPC w/ slurm (simple setup)
• working directory (???)

Workflow

• standardized structure
• standardized rule structure
  • benchmark and log files
• use tmp, shadow rules for temp output
• conda env. YAML files: update/clean
• config validation
  • schemas for config and sample table
• save config to output
• set working directory
• replace checkpoints with split/gather
  • consider whether splitting should be kept for all steps or not
  • seqkit: split2
  • todo: need to kno w the total number of seq.s
  • todo: signalp limit for number of seq.s?
• allow different step combinations
• output sub-folder per sample

# example for step a combination
workflows:
  - vir: true
  - amr: true
  - tox: false

Rules

• rule resources

Testing

• Use snakemake's unit test utility
• CI

GitLab

• templates for issues