{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://conformome.org/schema/molecule.schema.json",
  "title": "Molecule",
  "description": "One page per molecule: a protein (keyed on UniProt, isoform-aware), an RNA (RNAcentral URS, with its Rfam family), a DNA structure (sequence hash plus NAKB/PDB identifiers) or a complex, which links the pages of its components.",
  "type": "object",
  "required": [
    "molecule_id",
    "molecule_type",
    "identifiers",
    "name",
    "gene",
    "organism",
    "sequence_length",
    "sequence",
    "modifications",
    "domains",
    "ligands",
    "components",
    "provenance"
  ],
  "additionalProperties": false,
  "properties": {
    "molecule_id": {
      "$ref": "common.schema.json#/$defs/molecule_id"
    },
    "molecule_type": {
      "enum": [
        "protein",
        "rna",
        "dna",
        "complex"
      ]
    },
    "identifiers": {
      "type": "object",
      "required": [
        "uniprot",
        "isoform",
        "rnacentral",
        "rfam",
        "genbank",
        "nakb",
        "pdb"
      ],
      "additionalProperties": false,
      "properties": {
        "uniprot": {
          "type": [
            "string",
            "null"
          ],
          "pattern": "^([OPQ][0-9][A-Z0-9]{3}[0-9]|[A-NR-Z][0-9]([A-Z][A-Z0-9]{2}[0-9]){1,2})$"
        },
        "isoform": {
          "type": [
            "string",
            "null"
          ],
          "pattern": "^[A-Z0-9]{6,10}-[0-9]+$"
        },
        "rnacentral": {
          "type": [
            "string",
            "null"
          ],
          "pattern": "^URS[0-9A-F]{10}(_[0-9]+)?$"
        },
        "rfam": {
          "type": [
            "string",
            "null"
          ],
          "pattern": "^RF[0-9]{5}$"
        },
        "genbank": {
          "type": [
            "string",
            "null"
          ]
        },
        "nakb": {
          "type": [
            "string",
            "null"
          ]
        },
        "pdb": {
          "type": "array",
          "items": {
            "type": "string",
            "pattern": "^[0-9][A-Za-z0-9]{3}$"
          },
          "description": "PDB entries that define a DNA or RNA structure page (anchor entries)."
        }
      }
    },
    "name": {
      "type": "string",
      "minLength": 1
    },
    "gene": {
      "type": [
        "string",
        "null"
      ]
    },
    "organism": {
      "oneOf": [
        {
          "type": "null",
          "description": "Synthetic constructs and complexes across species."
        },
        {
          "type": "object",
          "required": [
            "name",
            "taxon_id"
          ],
          "additionalProperties": false,
          "properties": {
            "name": {
              "type": "string",
              "minLength": 1
            },
            "taxon_id": {
              "type": "integer",
              "minimum": 1
            }
          }
        }
      ]
    },
    "sequence_length": {
      "type": "integer",
      "minimum": 0,
      "description": "0 for complexes, whose sequences are on the component pages."
    },
    "sequence": {
      "type": "string",
      "description": "One-letter sequence. Protein: amino acids. RNA: ACGU (T allowed for DNA-RNA hybrids, N for unknown). DNA: ACGT. Complex: empty."
    },
    "modifications": {
      "type": "array",
      "description": "Post-translational modifications and modified nucleotides.",
      "items": {
        "type": "object",
        "required": [
          "position",
          "kind",
          "name",
          "source"
        ],
        "additionalProperties": false,
        "properties": {
          "position": {
            "type": "integer",
            "minimum": 1
          },
          "kind": {
            "enum": [
              "ptm",
              "modified_nucleotide",
              "other"
            ]
          },
          "name": {
            "type": "string",
            "minLength": 1,
            "description": "For example Phosphoserine, N6-methyladenosine (m6A), pseudouridine."
          },
          "source": {
            "type": "string",
            "minLength": 1
          }
        }
      }
    },
    "domains": {
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "name",
          "start",
          "end"
        ],
        "additionalProperties": false,
        "properties": {
          "name": {
            "type": "string"
          },
          "start": {
            "type": "integer",
            "minimum": 1
          },
          "end": {
            "type": "integer",
            "minimum": 1
          }
        }
      },
      "description": "Protein domains, or RNA elements (helices, aptamer, expression platform)."
    },
    "ligands": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "components": {
      "type": "array",
      "description": "Component molecules of a complex, each with its own page. Empty for other molecule types.",
      "items": {
        "type": "object",
        "required": [
          "molecule_id",
          "copies"
        ],
        "additionalProperties": false,
        "properties": {
          "molecule_id": {
            "$ref": "common.schema.json#/$defs/molecule_id"
          },
          "copies": {
            "type": [
              "integer",
              "null"
            ],
            "minimum": 1
          }
        }
      }
    },
    "provenance": {
      "$ref": "common.schema.json#/$defs/provenance"
    }
  },
  "allOf": [
    {
      "if": {
        "properties": {
          "molecule_type": {
            "const": "protein"
          }
        }
      },
      "then": {
        "properties": {
          "sequence": {
            "pattern": "^[A-Z]+$"
          },
          "identifiers": {
            "properties": {
              "uniprot": {
                "type": "string"
              }
            }
          },
          "components": {
            "maxItems": 0
          }
        }
      }
    },
    {
      "if": {
        "properties": {
          "molecule_type": {
            "const": "rna"
          }
        }
      },
      "then": {
        "properties": {
          "sequence": {
            "pattern": "^[ACGUTN]+$"
          },
          "identifiers": {
            "properties": {
              "rnacentral": {
                "type": "string"
              }
            }
          },
          "components": {
            "maxItems": 0
          }
        }
      }
    },
    {
      "if": {
        "properties": {
          "molecule_type": {
            "const": "dna"
          }
        }
      },
      "then": {
        "properties": {
          "sequence": {
            "pattern": "^[ACGTN]+$"
          },
          "components": {
            "maxItems": 0
          }
        }
      }
    },
    {
      "if": {
        "properties": {
          "molecule_type": {
            "const": "complex"
          }
        }
      },
      "then": {
        "properties": {
          "sequence": {
            "maxLength": 0
          },
          "components": {
            "minItems": 2
          }
        }
      }
    }
  ]
}