{
    "object_id": "https://biocomputeobject.org/GLY_001051/v-2.12.1",
    "spec_version": "1.3.0",
    "etag": "31e7d73677901094856ac124eaf6ec2c",
    "provenance_domain": {
        "name": "Human blood platelet Glycosylation Sites",
        "version": "v-2.12.1",
        "license": "https://creativecommons.org/licenses/by/4.0/",
        "created": "2024-08-01T12:27:42",
        "modified": "2024-08-06T13:58:24.773Z",
        "contributors": [
            {
                "name": "Jeet Vora",
                "affiliation": "The George Washington University",
                "email": "jeetvora@gwu.edu",
                "contribution": [
                    "createdBy",
                    "curatedBy"
                ]
            },
            {
                "name": "Karina Martinez",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "karinamartinez@gwu.edu"
            },
            {
                "name": "Robel Kahsay",
                "affiliation": "The George Washington University",
                "contribution": [
                    "createdBy"
                ],
                "email": "rykahsay@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"
            }
        ]
    },
    "usability_domain": [
        "The Human blood platelet Glycosylation Sites dataset contains glycosylation sites of thrombin-treated platelet proteins from Homo sapiens (taxid:9606). ESI-MS/MS and unbiased post-translational modification analysis of high-dose thrombin-treated releasate proteins, revealed O-glycosylation as a common modification of the secreted proteins and diverse O-glycan structures. A new form of domain-specific O-fucosylation was also detected within the elastin microfibril interface (EMI) domain of MMRN1, and was found to be important for it's secretion. The raw MS data can be found in the PRIDE repository under ID PXD045535.",
        "The dataset was generated collaboratively by labs at several institutes: The Heart Research Institute, Central Clinical School, and Charles Perkins Centre, at the University of Sydney in Australia, the Department of Biochemistry and Molecular Biology at the University of Georgia in the USA, and the Institute for Glyco-Core Research at the Nagoya University in Japan."
    ],
    "description_domain": {
        "keywords": [
            "o-glycosylation",
            "o-fucosylation",
            "o-mannosylation",
            "o-linked",
            "platelet",
            "c-linked"
        ],
        "pipeline_steps": [
            {
                "name": "make-proteoform-dataset.py",
                "step_number": 1,
                "input_list": [],
                "description": "The script maps the UniProtKB accesions to the UniProtKB Canonical accessions and also checks the position and amino acid of O-Gluc modified residue against UniProtKB Canonical sequences.",
                "prerequisite": [],
                "output_list": []
            }
        ]
    },
    "parametric_domain": [
        {
            "value": ""
        }
    ],
    "io_domain": {
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/human_proteoform_glycosylation_sites_platelet.csv",
                    "filename": "human_proteoform_glycosylation_sites_platelet.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/human_proteoform_glycosylation_sites_platelet.stat.csv",
                    "filename": "human_proteoform_glycosylation_sites_platelet.stat.csv"
                }
            }
        ],
        "input_subdomain": [
            {
                "uri": {
                    "filename": "human_proteoform_glycosylation_sites_platelet.csv",
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.12.1/compiled/human_proteoform_glycosylation_sites_platelet.csv"
                }
            }
        ]
    },
    "execution_domain": {
        "script": [
            {
                "uri": {
                    "uri": "https://github.com/glygener/glygen-backend-integration/blob/master/dataset-maker/make-proteoform-dataset.py",
                    "filename": "make-proteoform-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": "Homo sapiens",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Proteoform",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "O-Glucosylation Sites",
                        "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": "glycosylation_sites_"
                    }
                ]
            }
        }
    ],
    "error_domain": {}
}