{
    "object_id": "https://biocomputeobject.org/GLY_000378/v-2.5.1",
    "spec_version": "1.3.0",
    "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": "Hepatitis C virus (isolate H77)",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Protein",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "ViPR cross-references",
                        "category_name": "tag"
                    },
                    {
                        "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": "HCV1a UniProtKB Xref ViPR",
        "contributors": [
            {
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "name": "Jeet Vora",
                "email": "jeetvora@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"
            },
            {
                "name": "Luke Johnson",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "luke.johnson@email.gwu.edu"
            },
            {
                "name": "Urnisha Bhuiyan",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "sbhuiyan28@email.gwu.edu"
            },
            {
                "name": "Cyrus Au-Yeung",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "cyruschauyeung@gwu.edu"
            },
            {
                "name": "Kate Warner",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "k.warner1@gwu.edu"
            }
        ],
        "created": "2020-03-15T11:59:06.096223",
        "modified": "2024-05-17T11:23:58.322Z",
        "version": "v-2.5.1"
    },
    "description_domain": {
        "keywords": [
            "VIrus Pathogen Resource (ViPR)",
            "Hepaptitis C Virus (HCV)",
            "cross-reference"
        ],
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "step_number": 1,
                "input_list": [],
                "description": "The script loads the UniProtKB canonical accessions that are cross-referenced to VIrus Pathogen Resource (ViPR) database accessions/identifiers from the triplestore containing triples from nt file and extracts the data as a csv dataset",
                "prerequisite": [],
                "output_list": []
            }
        ]
    },
    "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/glygener/glygen-backend-integration/blob/master/dataset-maker/make-protein-dataset.py",
                    "filename": "make-protein-dataset.py"
                }
            }
        ]
    },
    "error_domain": {},
    "parametric_domain": [
        {
            "value": ""
        }
    ],
    "usability_domain": [
        "The HCV1a UniProtKB Xref Virus Pathogen Resource (ViPR) contains Hepatitis C virus (isolate H77) [taxid:63746] UniProtKB canonical accessions cross-referenced to Pfam database accessions/identifiers. If you use this dataset please provide proper attribution to EMBL-EBI-UniProtKB and GlyGen.",
        "The Virus Pathogen Database and Analysis Resource (ViPR) is an integrative and comprehensive publicly available database and analysis resource to search, analyze, visualize, save and share data for viral pathogens in the U.S. National Institute of Allergy and Infectious Diseases (NIAID) Category A-C Priority Pathogen lists for biodefense research, and other viral pathogens causing emerging/reemerging infectious diseases."
    ],
    "etag": "113544a21f869218f18000d3353ed72b",
    "io_domain": {
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/hcv1a_protein_xref_viruspathogenresource.csv",
                    "filename": "hcv1a_protein_xref_viruspathogenresource.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/hcv1a_protein_xref_viruspathogenresource.stat.csv",
                    "filename": "hcv1a_protein_xref_viruspathogenresource.stat.csv"
                }
            }
        ],
        "input_subdomain": [
            {
                "uri": {
                    "filename": "uniprot-proteome-hepatitis-c-virus-1a.nt",
                    "uri": "https:/data.glygen.org/ln2downloads/ebi/2024_04_08/uniprot-proteome-hepatitis-c-virus-1a.nt"
                }
            }
        ]
    }
}