{
    "object_id": "https://biocomputeobject.org/GLY_000448/v-1.11.2",
    "spec_version": "1.3.0",
    "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": "SARS coronavirus (SARS-CoV-1)",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Protein",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "Intact",
                        "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": "SARS-CoV1 UniProtKB Xref IntAct",
        "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": "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": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Karina Martinez",
                "email": "karinamartinez@gwu.edu"
            },
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Amanda Bell",
                "email": "amandab2140@gwu.edu"
            },
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Jonathon Keeney",
                "email": "keeneyjg@gwu.edu"
            },
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Millicent Quartey",
                "email": "mquartey@email.gwu.edu"
            }
        ],
        "created": "2020-04-03T12:16:43.851012",
        "modified": "2020-04-09T18:51:46.638840",
        "version": "v-1.11.2"
    },
    "description_domain": {
        "keywords": [
            "molecular interaction data",
            "protein-protein interaction",
            "IntAct",
            "cross-reference"
        ],
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "step_number": 1,
                "version": "",
                "input_list": [
                    {
                        "uri": "https://data.glygen.org/ln2data/releases/data/v-1.11.2/reviewed/sarscov1_protein_masterlist.csv",
                        "filename": "sarscov1_protein_masterlist.csv"
                    },
                    {
                        "uri": "https://data.glygen.org/ln2data/releases/data/v-1.11.2/source/uniprot-proteome-human-sars-coronavirus.nt",
                        "filename": "uniprot-proteome-human-sars-coronavirus.nt"
                    }
                ],
                "output_list": [
                    {
                        "uri": "https://data.glygen.org/ln2data/releases/data/v-1.11.2/reviewed/sarscov1_protein_xref_intact.csv",
                        "filename": "sarscov1_protein_xref_intact.csv"
                    }
                ],
                "prerequisite": [],
                "description": "The script loads the UniProtKB canonical accessions that are cross-referenced to IntAct database accessions/identifiers from the triplestore containing triples from nt file and extracts the data as a csv dataset"
            }
        ]
    },
    "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"
                }
            }
        ]
    },
    "usability_domain": [
        "The SARS coronavirus (SARS-CoV-1) UniProtKB Xref IntAct dataset contains sarscov1 [taxid:694009] UniProtKB canonical accessions cross-referenced to IntAct database accessions/identifiers. The dataset is derived from 2020-01 UniProtKB release. If you use this dataset please provide proper attribution to EMBL-EBI-UniProtKB and GlyGen",
        "IntAct provides a freely available, open source database system and analysis tools for molecular interaction data. All interactions are derived from literature curation or direct user submissions and are freely available. The IntAct Team also produce the Complex Portal."
    ],
    "etag": "c5935beb77d3558ce5fd73b368abcf49",
    "_id": {
        "$oid": "5e87616bf5c47b70c1fb8f8d"
    },
    "io_domain": {
        "input_subdomain": [
            {
                "uri": {
                    "uri": "https://data.glygen.org//ln2data/releases/data/v-1.11.2/reviewed/sarscov1_protein_masterlist.csv",
                    "filename": "sarscov1_protein_masterlist.csv"
                }
            },
            {
                "uri": {
                    "uri": "https://data.glygen.org//ln2downloads/ebi/2021_03/uniprot-proteome-sars-coronavirus.nt",
                    "filename": "uniprot-proteome-human-sars-coronavirus.nt"
                }
            }
        ],
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "http://data.glygen.org/ln2data/releases/data/v-1.11.2/reviewed/sarscov1_protein_xref_intact.csv",
                    "filename": "sarscov1_protein_xref_intact.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "http://data.glygen.org/ln2data/releases/data/v-1.11.2/reviewed/sarscov1_protein_xref_intact.stat.csv",
                    "filename": "sarscov1_protein_xref_intact.stat.csv"
                }
            }
        ]
    }
}