{
    "object_id": "https://biocomputeobject.org/GLY_001462/v-2.11.1",
    "spec_version": "1.3.0",
    "etag": "40898fae45ff83aedf2dd121629b078d",
    "provenance_domain": {
        "name": "Human Xref HPO",
        "version": "v-2.11.1",
        "license": "https://creativecommons.org/licenses/by/4.0/",
        "created": "2025-10-20T10:17:01",
        "modified": "2025-10-20T11:09:00.088Z",
        "contributors": [
            {
                "name": "Jeet Vora",
                "affiliation": "The George Washington University",
                "email": "jeetvora@gwu.edu",
                "contribution": [
                    "createdBy"
                ]
            },
            {
                "name": "Robel Kahsay",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "rykahsay@gwu.edu"
            },
            {
                "name": "Urnisha Bhuiyan",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "sbhuiyan28@email.gwu.edu"
            },
            {
                "name": "Kate Warner",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "k.warner1@gwu.edu"
            },
            {
                "name": "Mariia Kim",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "maria.kim@gwu.edu"
            }
        ]
    },
    "usability_domain": [
        "The Human Phenotype Ontology (HPO) xref dataset contains Homo sapiens [taxid:9606] UniProtKB canonical accessions cross-referenced to NCBI gene accessions/identifiers in the HPO database. If you use this dataset please provide proper attribution to GlyGen and HPO.",
        "The HPO database provides a standardized vocabulary of phenotypic abnormalities encountered in human disease."
    ],
    "description_domain": {
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "step_number": 1,
                "input_list": [],
                "description": "The script parses the HPO genes_to_disease csv file to extract the NCBI gene IDs in HPO and then maps the NCBI IDs to the UniProtKB Canonical accessions.",
                "prerequisite": [],
                "output_list": []
            }
        ],
        "keywords": [
            "HPO",
            "cross-reference",
            "phenotype ontology",
            "disease"
        ]
    },
    "parametric_domain": [
        {
            "value": ""
        }
    ],
    "io_domain": {
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.11.1/reviewed/human_protein_xref_hpo_gene.csv",
                    "filename": "human_protein_xref_hpo_gene.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.11.1/reviewed/human_protein_xref_hpo_gene.stat.csv",
                    "filename": "human_protein_xref_hpo_gene.stat.csv"
                }
            }
        ],
        "input_subdomain": [
            {
                "uri": {
                    "filename": "human_protein_xref_geneid.csv",
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.11.1/reviewed/human_protein_xref_geneid.csv"
                }
            },
            {
                "uri": {
                    "filename": "genes_to_disease.csv",
                    "uri": "https:/data.glygen.org/ln2downloads/hpo/2026_01_06/genes_to_disease.csv"
                }
            }
        ]
    },
    "execution_domain": {
        "script": [
            {
                "uri": {
                    "uri": "https://github.com/glygener/glygen-backend-integration/blob/master/dataset-maker/make-protein-dataset.py",
                    "filename": "make-protein-dataset.py"
                }
            }
        ],
        "script_driver": "manual",
        "software_prerequisites": [
            {
                "version": "2.7.5",
                "name": "Python",
                "uri": {
                    "uri": "https://www.python.org/download/releases/2.7.5/"
                }
            }
        ],
        "external_data_endpoints": [],
        "environment_variables": {}
    },
    "extension_domain": [
        {
            "extension_schema": "https://www.w3id.org/biocompute/extension_domain/1.2.0/dataset/dataset_extension.json",
            "dataset_extension": {
                "additional_license": {
                    "data_license": "https://creativecommons.org/licenses/by/4.0/",
                    "script_license": "https://www.gnu.org/licenses/gpl-3.0.en.html"
                },
                "dataset_categories": [
                    {
                        "category_value": "Homo sapiens",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Protein",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "HPO cross-references",
                        "category_name": "tag"
                    },
                    {
                        "category_value": "csv",
                        "category_name": "file_type"
                    },
                    {
                        "category_value": "Reviewed",
                        "category_name": "status"
                    },
                    {
                        "category_value": "internal",
                        "category_name": "scope"
                    }
                ]
            }
        }
    ],
    "error_domain": {}
}