nextflow_schema.json

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://raw.githubusercontent.com/dessimozlab/FastOMA/main/nextflow_schema.json",
  "title": "FastOMA Pipeline Parameters",
  "description": "Fast orthologous matrix analysis pipeline",
  "type": "object",
  "$defs": {
    "input_options": {
      "title": "Input options",
      "type": "object",
      "description": "Define where the pipeline should find input data ",
      "required": ["input"],
      "properties": {
        "input": {
          "type": "string",
          "description": "Input data source: local directory, archive file, or remote URL",
          "help": "Can be: (1) Path to a local directory containing proteome/ subfolder and species_tree.nwk file, (2) Path to a local archive file (.tar.gz, .tgz, .zip), or (3) HTTP/HTTPS URL to download an archive. Archives will be automatically extracted and cached.",
          "fa_icon": "fas fa-folder-open",
          "examples": [
            "/path/to/dataset/",
            "/path/to/dataset.tar.gz",
            "https://zenodo.org/records/12345/files/dataset.tar.gz"
          ]
        },
        "proteome_folder": {
          "type": "string",
          "format": "directory-path",
          "description": "Path to input directory containing the proteome files in fasta format",
          "help": "Override the default proteome folder location. Only used when input is a local directory. If not specified, defaults to input/proteome/.",
          "fa_icon": "fas fa-folder-open"
        },
        "hogmap_in": {
          "type": "string",
          "format": "directory-path",
          "description": "Path to input directory containing the hogmap files",
          "fa_icon": "fas fa-folder-open",
          "help": "If provided, the hogmap files will be used instead of computing them from scratch with OMAmer. This can save time if you have already computed the hogmaps for your proteomes."
        },
        "species_tree": {
          "type": "string",
          "format": "file-path",
          "description": "Path to species tree file in Newick format",
          "fa_icon": "fas fa-tree",
          "pattern": "^\\S+\\.(nhx|nh|nwk)",
          "help": "Override the default species tree location. Only used when input is a local directory. If not specified, defaults to input/species_tree.nwk."
        },
        "splice_folder": {
          "type": "string",
          "format": "directory-path",
          "description": "Path to input directory containing the splice files",
          "fa_icon": "fas fa-folder-open",
          "help": "If provided, FastOMA will use splice variant information to select representative isoforms for each gene. Only used when input is a local directory."
        },
        "omamer_db": {
          "type": "string",
          "format": "file-path",
          "description": "Path to OMAmer database against which the proteomes will be mapped",
          "fa_icon": "fas fa-database",
          "help": "If not provided, the default OMAmer database (LUCA) will be used.",
          "default": "https://omabrowser.org/All/LUCA.h5"
        },
        "test_data_cache": {
          "type": "string",
          "format": "directory-path",
          "description": "Path where (remote) input archives will be stored and permanently cached",
          "fa_ison": "fas fa-folder-open"
        },
        "input_folder": {
          "type": "string",
          "hidden": true,
          "format": "directory-path",
          "description": "DEPRECATED: Use --input instead",
          "help_text": "This parameter has been renamed to --input. Please update your command line or configuration files."
        }
      }
    },
    "output_options": {
      "title": "Output options",
      "type": "object",
      "description": "Define where the pipeline should store the output data",
      "required": ["output_folder"],
      "properties": {
        "output_folder": {
          "type": "string",
          "default": "Output",
          "format": "directory-path",
          "description": "The output directory where the results will be saved",
          "fa_icon": "fas fa-folder-open"
        },
        "hogmap_folder": {
          "type": "string",
          "format": "directory-path",
          "description": "The output directory where the omamer hogmap files will be saved",
          "help": "If not provided, the hogmap files will be saved in the output_folder/hogmap directory.",
          "fa_icon": "fas fa-folder-open",
          "default": "Output/hogmap"
        }
      }
    },
    "processing_options": {
      "title": "Processing options",
      "type": "object",
      "description": "Parameters for processing and analysis",
      "properties": {
        "fasta_header_id_transformer": {
          "type": "string",
          "default": "noop",
          "enum": ["noop", "UniProt"],
          "description": "FASTA header ID transformer method"
        },
        "filter_method": {
          "type": "string",
          "default": "col-row-threshold",
          "description": "MSA filtering method",
          "enum": ["col-row-threshold", "col-elbow-row-threshold", "trimal"]
        },
        "filter_gap_ratio_row": {
          "type": "number",
          "default": 0.3,
          "minimum": 0,
          "maximum": 1,
          "description": "Gap ratio threshold for row filtering"
        },
        "filter_gap_ratio_col": {
          "type": "number",
          "default": 0.5,
          "minimum": 0,
          "maximum": 1
        },
        "nr_repr_per_hog": {
          "type": "integer",
          "default": 5,
          "minimum": 1
        },
        "min_sequence_length": {
          "type": "integer",
          "default": 20
        }
      }
    },
    "output_control": {
      "title": "Output control",
      "type": "object",
      "description": "Control what outputs are saved",
      "properties": {
        "write_msas": {
          "type": "boolean",
          "description": "Write multiple sequence alignments"
        },
        "write_genetrees": {
          "type": "boolean",
          "description": "Write gene trees"
        },
        "debug_enabled": {
          "type": "boolean",
          "description": "Enable debug mode"
        },
        "msa_folder": {
          "type": "string",
          "default": "Output/msa"
        },
        "temp_output": {
          "type": "string",
          "default": "Output/temp_output"
        },
        "genetrees_folder": {
          "type": "string",
          "default": "Output/genetrees"
        },
        "force_pairwise_ortholog_generation": {
          "type": "boolean",
          "description": "Force generation of pairwise orthologs even for large datasets (may be slow)",
          "help": "By default, pairwise orthologs are not generated for datasets with more than 25 species to avoid long runtimes. Enable this option to override this behavior."
        }
      }
    },
    "resource_options": {
      "title": "Resource options",
      "type": "object",
      "description": "",
      "default": "",
      "properties": {
        "time_multiplier": {
          "type": "integer",
          "default": 1
        },
        "memory_multiplier": {
          "type": "integer",
          "default": 1
        }
      }
    },
    "generic_options": {
      "title": "Generic options",
      "type": "object",
      "description": "Less common options for the pipeline",
      "properties": {
        "help": {
          "type": "boolean",
          "description": "Display help text",
          "fa_icon": "fas fa-question-circle",
          "hidden": true
        },
        "report": {
          "type": "boolean",
          "description": "Generate nextflow reports"
        },
        "statsdir": {
          "type": "string",
          "fa_icon": "fas fa-folder-open",
          "default": "Output/stats"
        },
        "trace_report_suffix": {
          "type": "string",
          "fa_icon": "far calendar",
          "description": "Suffix to add to the trace report filename. Default is the date and time in the format yyyy-MM-dd_HH-mm-ss.",
          "hidden": true
        },
        "custom_config_version": {
          "type": "string",
          "default": "master"
        },
        "custom_config_base": {
          "type": "string",
          "default": "https://raw.githubusercontent.com/nf-core/configs/master"
        },
        "container_name": {
          "type": "string",
          "default": "dessimozlab/fastoma"
        },
        "container_version": {
          "type": "string",
          "default": "dev"
        }
      }
    }
  },
  "allOf": [
    {
      "$ref": "#/$defs/input_options"
    },
    {
      "$ref": "#/$defs/output_options"
    },
    {
      "$ref": "#/$defs/processing_options"
    },
    {
      "$ref": "#/$defs/output_control"
    },
    {
      "$ref": "#/$defs/resource_options"
    },
    {
      "$ref": "#/$defs/generic_options"
    }
  ]
}