{
    "io_domain": {
        "input_subdomain": [
            {
                "uri": {
                    "uri": "https://data.glygen.org/ln2data/releases/data/v-1.4.5/source/uniprot-proteome-9606-homo-sapiens-2019-05.nt",
                    "filename": "uniprot-proteome-9606-homo-sapiens-2019-05.nt"
                }
            },
            {
                "uri": {
                    "uri": "https://data.glygen.org/ln2data/releases/data/v-1.4.5/compiled/human_protein_glyco_disease.csv",
                    "filename": "human_protein_glyco_disease.csv"
                }
            },
            {
                "uri": {
                    "uri": "https://data.glygen.org/ln2data/releases/data/v-1.4.5/reviewed/human_protein_idmapping.csv",
                    "filename": "human_protein_idmapping.csv"
                }
            }
        ],
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "http://data.glygen.org/ln2data/releases/data/v-1.4.5/reviewed/human_protein_disease.csv",
                    "filename": "human_protein_disease.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "http://data.glygen.org/ln2data/releases/data/v-1.4.5/reviewed/human_protein_disease.stat.csv",
                    "filename": "human_protein_disease.stat.csv"
                }
            }
        ]
    },
    "spec_version": "1.3.0",
    "object_id": "https://biocomputeobject.org/GLY_000084/v-1.4.5",
    "extension_domain": [
        {
            "extension_schema": "http://www.w3id.org/biocompute/extension_domain/1.1.0/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": "Diseases",
                        "category_name": "tags"
                    },
                    {
                        "category_value": "csv",
                        "category_name": "file_type"
                    },
                    {
                        "category_value": "Reviewed",
                        "category_name": "status"
                    },
                    {
                        "category_value": "internal",
                        "category_name": "scope"
                    }
                ]
            }
        }
    ],
    "provenance_domain": {
        "license": "https://creativecommons.org/licenses/by/4.0/",
        "name": "Human Diseases",
        "contributors": [
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Jeet Vora",
                "email": "jeetvora@gwu.edu"
            },
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Rahi Navelkar",
                "email": "rsn13@gwu.edu"
            },
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Robel Kahsay",
                "email": "rykahsay@gwu.edu"
            },
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Xavier Holmes",
                "email": "xavierh@email.gwu.edu"
            },
            {
                "affiliation": "EMBL-EBI-UniProt",
                "contribution": [
                    "contributedBy"
                ],
                "name": "Maria Martin",
                "email": "martin@ebi.ac.uk"
            },
            {
                "affiliation": "EMBL-EBI-UniProt",
                "contribution": [
                    "contributedBy"
                ],
                "name": "Preethi Vasudev",
                "email": "pvasudev@ebi.ac.uk"
            },
            {
                "affiliation": "Monarch Initiative - Oregon Health & Science University",
                "contribution": [
                    "contributedBy"
                ],
                "name": "Jean-Philippe Gourdine",
                "email": "gourdine@ohsu.edu"
            },
            {
                "affiliation": "Monarch Initiative - Oregon Health & Science University",
                "contribution": [
                    "contributedBy"
                ],
                "name": "Melissa Haendel",
                "email": "haendel@ohsu.edu"
            }
        ],
        "created": "2019-01-23T17:21:05.953432",
        "modified": "2019-09-12T14:43:38.058730",
        "version": "v-1.4.5"
    },
    "description_domain": {
        "keywords": [
            "disease",
            "HPO",
            "OMIM",
            "Congenital Disorders of Glycosylation (CDG)",
            "Monarch Initiative"
        ],
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "step_number": 1,
                "version": "",
                "input_list": [
                    {
                        "uri": "https://data.glygen.org/ln2data/releases/data/v-1.4.5/reviewed/human_protein_idmapping.csv",
                        "filename": "human_protein_idmapping.csv"
                    },
                    {
                        "uri": "https://data.glygen.org/ln2data/releases/data/v-1.4.5/source/uniprot-proteome-9606-homo-sapiens-2019-05.nt",
                        "filename": "uniprot-proteome-9606-homo-sapiens-2019-05.nt"
                    }
                ],
                "output_list": [
                    {
                        "uri": "https://data.glygen.org/ln2data/releases/data/v-1.4.5/compiled/human_protein_glyco_disease.csv",
                        "filename": "human_protein_glyco_disease.csv"
                    }
                ],
                "prerequisite": [],
                "description": "The script extracts the OMIM phenotypes for UniProtKB accessions and disease information from human_glyco-phenotype_disease_curation_list_v12-18-74.csv. The script maps the UniProtKB accessions to the UniprotKB canonical accessions."
            }
        ]
    },
    "execution_domain": {
        "external_data_endpoints": [],
        "environment_variables": {},
        "script_driver": "manual",
        "software_prerequisites": [
            {
                "version": "2.7.5",
                "name": "Python",
                "uri": {
                    "uri": "https://www.python.org/download/releases/2.7.5/"
                }
            }
        ],
        "script": [
            {
                "uri": {
                    "uri": "https://github.com/GW-HIVE/glygen-backend-integration/blob/master/pipeline/integrator/make-protein-dataset.py",
                    "filename": "make-protein-dataset.py"
                }
            }
        ]
    },
    "error_domain": {
        "empirical_error": {},
        "algorithmic_error": {},
        "emperical_error": {}
    },
    "parametric_domain": [
        {
            "step": "",
            "value": "",
            "param": ""
        }
    ],
    "usability_domain": [
        "The Human Diseases dataset contains list of human [taxid:9606] diseases as described by OMIM and Monarch Initiave database ",
        "Online Mendelian Inheritance in Man (OMIM) is a continuously updated catalog of human genes and genetic disorders and traits, with a particular focus on the gene-phenotype relationship.",
        "The Monarch Initiative is focused primarily on phenotype-related resources and brings in data associated with those phenotypes so that the users can begin to make connections among other biological entities of interest.",
        "If you are using this dataset please give proper attribution to OMIM, Monarch Initiative, EMBL-EBI-UniProt and GlyGen"
    ],
    "etag": "",
    "_id": {
        "$oid": "5c48e8d1f5c47b65ae4b5d3f"
    }
}
