{
    "object_id": "https://biocomputeobject.org/GLY_001681/v-2.12.1",
    "spec_version": "1.3.0",
    "etag": "12400600acad8e66da927754bbce3120",
    "provenance_domain": {
        "name": "SARS-CoV2 Binding Sites Citations (UniProtKB)",
        "version": "v-2.12.1",
        "license": "https://creativecommons.org/licenses/by/4.0/",
        "modified": "2026-08-18T21:04:47.980Z",
        "contributors": [
            {
                "name": "Jeet Vora",
                "affiliation": "The George Washington University",
                "email": "jeetvora@gwu.edu",
                "contribution": [
                    "createdBy"
                ]
            },
            {
                "name": "Robel Kahsay",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "rykahsay@gwu.edu"
            },
            {
                "name": "Urnisha Bhuiyan",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "sbhuiyan28@email.gwu.edu"
            },
            {
                "name": "Kate Warner",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "k.warner1@gwu.edu"
            }
        ]
    },
    "usability_domain": [
        "The SARS coronavirus (SARS-CoV-2 or 2019-nCov) binding sites citations (UniProtKB) dataset contains Severe acute respiratory syndrome coronavirus 2 [taxid:2697049] publication information for UniProtKB canonical protein binding sites provided in GlyGen. The dataset includes PMID, journal name, date and author and title. If you use this dataset please give proper attribution to UniProtKB, ChEBI and GlyGen."
    ],
    "description_domain": {
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py",
                "step_number": 1,
                "input_list": [],
                "description": "The script extracts the PMID from the Binding_Site_Annotation triple then accesses the NCBI ENTREZ infrastructure to retrieve the publication information of the given PMID from UniProtKB.",
                "prerequisite": [],
                "output_list": []
            }
        ],
        "keywords": [
            "binding sites",
            "ligands",
            "metal binding sites",
            "calcium binding sites",
            "citations"
        ]
    },
    "io_domain": {
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/sarscov2_protein_citations_binding_sites_uniprotkb.csv",
                    "filename": "sarscov2_protein_citations_binding_sites_uniprotkb.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/sarscov2_protein_citations_binding_sites_uniprotkb.stat.csv",
                    "filename": "sarscov2_protein_citations_binding_sites_uniprotkb.stat.csv"
                }
            }
        ],
        "input_subdomain": [
            {
                "uri": {
                    "filename": "sarscov2_protein_binding_sites_uniprotkb.csv",
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/sarscov2_protein_binding_sites_uniprotkb.csv"
                }
            }
        ]
    },
    "execution_domain": {
        "script": [
            {
                "uri": {
                    "uri": "https://github.com/glygener/glygen-backend-integration/blob/master/dataset-maker/make-protein-dataset.py",
                    "filename": "make-protein-dataset.py"
                }
            }
        ],
        "script_driver": "manual",
        "software_prerequisites": [
            {
                "version": "2.7.5",
                "name": "Python",
                "uri": {
                    "uri": "https://www.python.org/download/releases/2.7.5/"
                }
            }
        ],
        "external_data_endpoints": [],
        "environment_variables": {}
    },
    "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": "Severe acute respiratory syndrome coronavirus 2",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Protein",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "Binding sites citations",
                        "category_name": "tag"
                    },
                    {
                        "category_value": "csv",
                        "category_name": "file_type"
                    },
                    {
                        "category_value": "Reviewed",
                        "category_name": "status"
                    },
                    {
                        "category_value": "internal",
                        "category_name": "scope"
                    },
                    {
                        "category_name": "file_category",
                        "category_value": "citations"
                    }
                ]
            }
        }
    ],
    "error_domain": {}
}