{
    "object_id": "https://biocomputeobject.org/GLY_000951/v-2.7.1",
    "etag": "09206b2bcf5e933aaa7d585190eed133",
    "spec_version": "1.3.0",
    "provenance_domain": {
        "name": "Pig UniProtKB Xref RefSeq",
        "version": "v-2.7.1",
        "license": "https://creativecommons.org/licenses/by/4.0/",
        "created": "",
        "modified": "2024-05-17T11:24:19.596Z",
        "derived_from": [],
        "contributors": [
            {
                "name": "Jeet Vora",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "jeetvora@gwu.edu"
            },
            {
                "name": "Robel Kahsay",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "rykahsay@gwu.edu"
            },
            {
                "name": "Maria Martin",
                "affiliation": "EMBL-EBI-UniProt",
                "contribution": [
                    "contributedBy"
                ],
                "email": "martin@ebi.ac.uk"
            },
            {
                "name": "Preethi Vasudev",
                "affiliation": "EMBL-EBI-UniProt",
                "contribution": [
                    "contributedBy"
                ],
                "email": "pvasudev@ebi.ac.uk"
            },
            {
                "name": "Karina Martinez",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "karinamartinez@gwu.edu"
            },
            {
                "name": "Jingyue Wu",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "jingyue.wu@email.gwu.edu"
            },
            {
                "name": "Luke Johnson",
                "affiliation": "The George Washington University",
                "email": "luke.johnson@email.gwu.edu",
                "contribution": [
                    "createdBy"
                ],
                "orcid": ""
            },
            {
                "name": "Kate Warner",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "k.warner1@gwu.edu"
            }
        ],
        "review": []
    },
    "usability_domain": [
        "The Pig UniprotKB Xref RefSeq contains Sus scrofa [taxid:9823] 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. "
    ],
    "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": "Homo sapiens",
                        "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"
                    }
                ]
            }
        }
    ],
    "description_domain": {
        "keywords": [
            "protein accessions",
            "RefSeq",
            "cross-reference"
        ],
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "version": "",
                "step_number": 1,
                "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"
                }
            }
        ]
    },
    "parametric_domain": [
        {
            "step": "",
            "param": "",
            "value": ""
        }
    ],
    "io_domain": {
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.7.1/reviewed/pig_protein_xref_refseq.csv",
                    "filename": "pig_protein_xref_refseq.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.7.1/reviewed/pig_protein_xref_refseq.stat.csv",
                    "filename": "pig_protein_xref_refseq.stat.csv"
                }
            }
        ],
        "input_subdomain": [
            {
                "uri": {
                    "filename": "uniprot-proteome-sus-scrofa.nt",
                    "uri": "https:/data.glygen.org/ln2downloads/ebi/2024_09_19/uniprot-proteome-sus-scrofa.nt"
                }
            }
        ]
    },
    "error_domain": {
        "empirical_error": {},
        "algorithmic_error": {}
    }
}