{
    "io_domain": {
        "input_subdomain": [
            {
                "uri": {
                    "uri": "https://data.glygen.org/ln2wwwdata/source/uniprot-proteome-hepatitis-c-virus-1b.nt", 
                    "filename": "uniprot-proteome-hepatitis-c-virus-1b.nt"
                }
            }
        ], 
        "output_subdomain": [
            {
                "mediatype": "csv", 
                "uri": {
                    "uri": "http://data.glygen.org/ln2wwwdata/reviewed/hcv1b_protein_xref_viruspathogenresource.csv", 
                    "filename": "hcv1b_protein_xref_viruspathogenresource.csv"
                }
            }, 
            {
                "mediatype": "csv", 
                "uri": {
                    "uri": "http://data.glygen.org/ln2wwwdata/reviewed/hcv1b_protein_xref_viruspathogenresource.stat.csv", 
                    "filename": "hcv1b_protein_xref_viruspathogenresource.stat.csv"
                }
            }
        ]
    }, 
    "spec_version": "1.3.0", 
    "object_id": "https://biocomputeobject.org/GLY_000379/1.7.10", 
    "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": "Hepatitis C virus (genotype 1b, isolate Japanese)", 
                        "category_name": "species"
                    }, 
                    {
                        "category_value": "Protein", 
                        "category_name": "molecule"
                    }, 
                    {
                        "category_value": "ViPR cross-references", 
                        "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": "HCV1b 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": "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"
            }
        ], 
        "created": "2020-03-15T12:01:51.641730", 
        "modified": "2020-04-09T18:10:47.193099", 
        "version": "1.7.10"
    }, 
    "description_domain": {
        "keywords": [
            "VIrus Pathogen Resource (ViPR)", 
            "Hepaptitis C Virus (HCV)", 
            "cross-reference"
        ], 
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py", 
                "step_number": 1, 
                "version": "", 
                "input_list": [
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/reviewed/hcv1b_protein_masterlist.csv", 
                        "filename": "hcv1b_protein_masterlist.csv"
                    }, 
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/source/uniprot-proteome-hepatitis-c-virus-1b.nt", 
                        "filename": "uniprot-proteome-hepatitis-c-virus-1b.nt"
                    }
                ], 
                "output_list": [
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/reviewed/hcv1b_protein_xref_viruspathogenresource.csv", 
                        "filename": "hcv1b_protein_xref_viruspathogenresource.csv"
                    }
                ], 
                "prerequisite": [], 
                "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"
            }
        ]
    }, 
    "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 Hepatitis C virus (genotype 1b, isolate Japanese) UniProtKB Xref Virus Pathogen Resource contains hcv1b [taxid:11116] UniProtKB canonical accessions cross-referenced to Pfam database accessions/identifiers. The dataset is derived from 2019-09 UniProtKB release. 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": "4833e415abb9ffce089d0a67fc6e2dc4", 
    "_id": {
        "$oid": "5e6e516ff5c47b55a95d8e6b"
    }
}