name: "gecco_run"
description: GECCO is a fast and scalable method for identifying putative novel Biosynthetic Gene Clusters (BGCs) in genomic and metagenomic data using Conditional Random Fields (CRFs).
keywords:
  - bgc
  - detection
  - metagenomics
  - contigs
tools:
  - "gecco":
      description: "Biosynthetic Gene Cluster prediction with Conditional Random Fields."
      homepage: "https://gecco.embl.de"
      documentation: "https://gecco.embl.de"
      tool_dev_url: "https://github.com/zellerlab/GECCO"
      doi: "10.1101/2021.05.03.442509"
      licence: "['GPL v3']"

input:
  - meta:
      type: map
      description: |
        Groovy Map containing sample information
        e.g. [ id:'test', single_end:false ]
  - input:
      type: file
      description: A genomic file containing one or more sequences as input. Input type is any supported by Biopython (fasta, gbk, etc.)
      pattern: "*"
  - hmm:
      file: file
      description: Alternative HMM file(s) to use in HMMER format
      pattern: "*.hmm"
  - model_dir:
      file: directory
      description: Path to an alternative CRF (Conditional Random Fields) module to use

output:
  - meta:
      type: map
      description: |
        Groovy Map containing sample information
        e.g. [ id:'test', single_end:false ]
  - versions:
      type: file
      description: File containing software versions
      pattern: "versions.yml"
  - genes:
      type: file
      description: TSV file containing detected/predicted genes with BGC probability scores. Will not be generated if no hits are found.
      pattern: "*.genes.tsv"
  - features:
      type: file
      description: TSV file containing identified domains
      pattern: "*.features.tsv"
  - clusters:
      type: file
      description: TSV file containing coordinates of predicted clusters and BGC types.  Will not be generated if no hits are found.
      pattern: "*.clusters.tsv"
  - gbk:
      type: file
      Description: Per cluster GenBank file (if found) containing sequence with annotations. Will not be generated if no hits are found.
      pattern: "*.gbk"
  - json:
      type: file
      description: AntiSMASH v6 sideload JSON file (if --antismash-sideload) supplied. Will not be generated if no hits are found.
      pattern: "*.gbk"

authors:
  - "@jfy133"