{
    "object_id": "https://biocomputeobject.org/GLY_001251/v-2.9.1",
    "spec_version": "1.3.0",
    "etag": "a2b5dcc660965116054f7467f3948ebb",
    "provenance_domain": {
        "name": "Mouse Glycosylation Sites (GlyCosmos)",
        "version": "v-2.9.1",
        "license": "https://creativecommons.org/licenses/by/4.0/",
        "created": "2022-08-16T20:48:29.540934",
        "modified": "2025-08-03T22:24:30.214Z",
        "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 GlyCosmos Mouse Glycosylation Sites dataset contains information on Asn (N)-glycosylated proteins and their glycosylation sites in *Mus musculus* [taxid:10090]. This dataset originates from GlycoProtDB (https://acgg.asia/db/gpdb/) and is made available to GlyGen through GlyCosmos (https://glycosmos.org/). The identification of glycoproteins in mouse models was carried out using LC/MS techniques as part of the Structural Glycomics Project, supported by NEDO (New Energy and Industrial Technology Development Organization) and the RIO Database Project by AIST. The development of the database was further supported by MEXT\u2019s Integrated Database Project and the Program for Coordination Toward Integration of Related Databases, led by the Japan Science and Technology Agency (JST)."
    ],
    "description_domain": {
        "pipeline_steps": [
            {
                "name": "make-proteoform-dataset.py",
                "step_number": 1,
                "description": "The script maps the UniProtKB accesions to the UniProtKB Canonical accessions and also checks the position and amino acid of O-GlcNAc modified residue against UniProtKB Canonical sequences. The script that processes this dataset matches the amino acid and position with the latest release fasta file as a QC check to eliminate the incorrect entries. The script also adds the -5 and +5 amino acids from the modified residue in the peptide field.",
                "input_list": [],
                "prerequisite": [],
                "output_list": []
            }
        ],
        "keywords": [
            "Glycosylation Sites",
            "GlyCosmos",
            "GlycoProtDB"
        ]
    },
    "io_domain": {
        "output_subdomain": [
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/mouse_proteoform_glycosylation_sites_glycosmos.csv",
                    "filename": "mouse_proteoform_glycosylation_sites_glycosmos.csv"
                }
            },
            {
                "mediatype": "csv",
                "uri": {
                    "uri": "https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/mouse_proteoform_glycosylation_sites_glycosmos.stat.csv",
                    "filename": "mouse_proteoform_glycosylation_sites_glycosmos.stat.csv"
                }
            }
        ],
        "input_subdomain": [
            {
                "uri": {
                    "filename": "filtered_glycoproteins_glycan_list.csv",
                    "uri": "https:/data.glygen.org/ln2downloads/glycosmos/2025_06_23/filtered_glycoproteins_glycan_list.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": "Mus musculus",
                        "category_name": "species"
                    },
                    {
                        "category_value": "Proteoform",
                        "category_name": "molecule"
                    },
                    {
                        "category_value": "Glycosylation Sites (GlyCosmos)",
                        "category_name": "tag"
                    },
                    {
                        "category_value": "csv",
                        "category_name": "file_type"
                    },
                    {
                        "category_value": "Reviewed",
                        "category_name": "status"
                    },
                    {
                        "category_value": "internal",
                        "category_name": "scope"
                    }
                ]
            }
        }
    ],
    "error_domain": {}
}