{
    "object_id": "https://biocomputeobject.org/GLY_000128/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": "Mus musculus",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Protein",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "Refseq 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": "Mouse UniProtKB Xref RefSeq",
        "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": "2019-03-15T16:23:15.480143",
        "modified": "2024-05-17T11:23:52.040Z",
        "version": "v-2.5.1"
    },
    "description_domain": {
        "keywords": [
            "protein accessions",
            "RefSeq",
            "cross-reference"
        ],
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "step_number": 1,
                "version": "",
                "input_list": [],
                "output_list": [],
                "prerequisite": [],
                "description": "The script loads the UniProtKB canonical accessions that are cross-referenced to NCBI RefSeq 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/glygener/glygen-backend-integration/blob/master/dataset-maker/make-protein-dataset.py",
                    "filename": "make-protein-dataset.py"
                }
            }
        ]
    },
    "error_domain": {
        "empirical_error": {},
        "algorithmic_error": {}
    },
    "parametric_domain": [
        {
            "step": "",
            "param": "",
            "value": ""
        }
    ],
    "usability_domain": [
        "The Mouse UniprotKB Xref RefSeq contains Mus musculus [taxid:10090] UniProtKB canonical accessions cross-referenced to RefSeq database accessions/identifiers. The script extracts one-to-one UniProtKB isoform ac to RefSeq ac and also UniProtKB ac to RefSeq ac where there is no UniProtKB isoforms other than the canonical ac. The dataset is contributed by EMBL-EBI-UniProtKB and processed by GlyGen. If you use this dataset please provide proper attribution to EMBL-EBI-UniProtKB and GlyGen. ",
        "NCBI RefSeq is a comprehensive, integrated, non-redundant, well-annotated set of reference sequences including genomic, transcript, and protein. "
    ],
    "etag": "da0064c9c09e5420d1a33966327d0fd7",
    "io_domain": {
        "input_subdomain": [
            {
                "uri": {
                    "filename": "uniprot-proteome-mus-musculus.nt",
                    "uri": "https:/data.glygen.org/ln2downloads/ebi/2024_04_08/uniprot-proteome-mus-musculus.nt"
                }
            }
        ],
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/mouse_protein_xref_refseq.csv",
                    "filename": "mouse_protein_xref_refseq.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/mouse_protein_xref_refseq.stat.csv",
                    "filename": "mouse_protein_xref_refseq.stat.csv"
                }
            }
        ]
    }
}