{
    "object_id": "https://biocomputeobject.org/GLY_000395/1.8.25", 
    "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": "Homo sapiens", 
                        "category_name": "species"
                    }, 
                    {
                        "category_value": "Protein", 
                        "category_name": "molecule"
                    }, 
                    {
                        "category_value": "Protein Submitted Names", 
                        "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 Protein Submitted Names (UniProtKB)", 
        "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": "2020-03-15T13:38:29.540245", 
        "modified": "2020-04-10T11:49:36.985464", 
        "version": "1.8.25"
    }, 
    "description_domain": {
        "keywords": [
            "protein submitted name", 
            "protein name", 
            "UniProtKB"
        ], 
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py", 
                "step_number": 1, 
                "version": "", 
                "input_list": [
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/reviewed/human_protein_masterlist.csv", 
                        "filename": "human_protein_masterlist.csv"
                    }, 
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/source/uniprot-proteome-homo-sapiens.nt", 
                        "filename": "uniprot-proteome-homo-sapiens.nt"
                    }
                ], 
                "output_list": [
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/reviewed/human_protein_submittednames.csv", 
                        "filename": "human_protein_submittednames.csv"
                    }
                ], 
                "prerequisite": [], 
                "description": "The script loads the UniProtKB submitted names for protein accessions from the triplestore containing triples from nt file and extracts the recommended names into csv dataset file"
            }
        ]
    }, 
    "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/pipeline/extractor/extract-from-endpoint.py", 
                    "filename": "extract-from-endpoint.py"
                }
            }
        ]
    }, 
    "error_domain": {
        "empirical_error": {}, 
        "algorithmic_error": {}, 
        "emperical_error": {}
    }, 
    "parametric_domain": [
        {
            "step": "", 
            "param": "", 
            "value": ""
        }
    ], 
    "usability_domain": [
        "The Human Protein Submitted Names (UniProtKB) dataset contains UniProtKB submitted names for human [taxid:9606] UniProtKB/TrEMBL proteins accessions. In UniProt, UniProtKB/TrEMBL entries usually have 'Submitted name' which may later be improved by automatic annotation procedures but if not, it remains as provided by the submitter until the entry is manually annotated and integrated to UniProtKB/Swiss-Prot. The dataset is derived from 2019-09 UniProtKB release. If you use this dataset please provide proper attribution to EMBL-EBI-UniProtKB and GlyGen"
    ], 
    "etag": "9ec5b568fc6336197464c2710b3b8153", 
    "_id": {
        "$oid": "5e6e6815f5c47b5ec035001f"
    }, 
    "io_domain": {
        "input_subdomain": [
            {
                "uri": {
                    "uri": "https://data.glygen.org//ln2downloads/ebi/2020_12/uniprot-proteome-homo-sapiens.nt", 
                    "filename": "uniprot-proteome-homo-sapiens.nt"
                }
            }, 
            {
                "uri": {
                    "uri": "https://data.glygen.org//ln2releases/data/v-1.8.25/reviewed/human_protein_masterlist.csv", 
                    "filename": "human_protein_masterlist.csv"
                }
            }
        ], 
        "output_subdomain": [
            {
                "mediatype": "csv", 
                "uri": {
                    "uri": "http://data.glygen.org/ln2wwwdata/reviewed/human_protein_submittednames.csv", 
                    "filename": "human_protein_submittednames.csv"
                }
            }, 
            {
                "mediatype": "csv", 
                "uri": {
                    "uri": "http://data.glygen.org/ln2wwwdata/reviewed/human_protein_submittednames.stat.csv", 
                    "filename": "human_protein_submittednames.stat.csv"
                }
            }
        ]
    }
}