json

Files

• hetionet-v1.0.json.bz2 is a bzip2-compressed JSON text file containing the top-level objects described below.

• hetionet-v1.0-metagraph.json is a JSON text file containing only the objects describing the type graph (called a metagraph or network schema). These are metanode_kinds, metaedge_tuples, and kind_to_abbrev.

JSON hetnet format

The JSON format has the following top-level objects:

• metanode_kinds: an array of node types. Example:

"metanode_kinds": [
  "Gene",
  "Molecular Function"
]

• metaedge_tuples: an array of edge types. Each edge type is represented as an array with positional values corresponding to the source node type, target edge type, edge type, and directionality. Example:

"metaedge_tuples": [
  [
    "Gene",
    "Gene",
    "regulates",
    "forward"
  ],
  [
    "Gene",
    "Molecular Function",
    "participates",
    "both"
  ]
]

• kind_to_abbrev: an object mapping node and edge types to their abbreviations. Note that we include this abbreviation mapping, but consider an optional component to the JSON format. Example:

"kind_to_abbrev": {
  "Molecular Function": "MF",
  "Gene": "G",
  "participates": "p"
}

• nodes: an array of nodes. Each node is represented as an object with pairs for the node type (kind), identifier, name, and properties (data). Node identifiers are guaranteed to be unique within each node type. Example:

"nodes": [
  {
    "kind": "Molecular Function",
    "identifier": "GO:0031753",
    "name": "endothelial differentiation G-protein coupled receptor binding",
    "data": {
      "source": "Gene Ontology",
      "license": "CC BY 4.0",
      "url": "http://purl.obolibrary.org/obo/GO_0031753"
    }
  },
  {
    "kind": "Gene",
    "identifier": 5345,
    "name": "SERPINF2",
    "data": {
      "description": "serpin peptidase inhibitor, clade F (alpha-2 antiplasmin, pigment epithelium derived factor), member 2",
      "source": "Entrez Gene",
      "license": "CC0 1.0",
      "url": "http://identifiers.org/ncbigene/5345",
      "chromosome": "17"
    }
  }
]

• edges: an array of edges. Each edge is represented as an object with pairs for the source node (source_id), target node (target_id), edge type (kind), directionality, and propertier (data). Example:

"edges": [
  {
    "source_id": [
      "Gene",
      2160
    ],
    "target_id": [
      "Molecular Function",
      "GO:0008201"
    ],
    "kind": "participates",
    "direction": "both",
    "data": {
      "source": "NCBI gene2go",
      "license": "CC BY 4.0",
      "unbiased": false
    }
  },
  {
    "source_id": [
      "Compound",
      "DB01106"
    ],
    "target_id": [
      "Gene",
      4931
    ],
    "kind": "downregulates",
    "direction": "both",
    "data": {
      "source": "LINCS L1000",
      "z_score": -4.156,
      "method": "measured",
      "unbiased": true
    }
  }
]

JSON to CSV

To expedite data loading through CSV format, execute the following Python script using Python 3:

python json_to_csv.py

Note: The existing data in the CSV folder has already been produced by this script. No need to run it again unless the JSON data is updated.