{
    "io_domain": {
        "input_subdomain": [
            {
                "uri": {
                    "uri": "https://data.glygen.org/ln2wwwdata/source/congenital_disorders_of_glycosylation.tsv", 
                    "filename": "congenital_disorders_of_glycosylation.tsv"
                }
            }, 
            {
                "uri": {
                    "uri": "https://data.glygen.org/ln2wwwdata/reviwed/protein_domap.csv", 
                    "filename": "protein_domap.csv"
                }
            }
        ], 
        "output_subdomain": [
            {
                "mediatype": "csv", 
                "uri": {
                    "uri": "http://data.glygen.org/ln2wwwdata/reviewed/human_protein_genomics_england_disease.csv", 
                    "filename": "human_protein_disease_genomics_england.csv"
                }
            }, 
            {
                "mediatype": "csv", 
                "uri": {
                    "uri": "http://data.glygen.org/ln2wwwdata/reviewed/human_protein_genomics_england_disease.stat.csv", 
                    "filename": "human_protein_genomics_england_disease.stat.csv"
                }
            }
        ]
    }, 
    "spec_version": "1.3.0", 
    "object_id": "https://biocomputeobject.org/GLY_000308/1.8.24", 
    "extension_domain": [
        {
            "extension_schema": "http://www.w3id.org/biocompute/extension_domain/1.1.0/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": "Disease", 
                        "category_name": "tags"
                    }, 
                    {
                        "category_value": "Reviewed", 
                        "category_name": "status"
                    }, 
                    {
                        "category_value": "csv", 
                        "category_name": "file_type"
                    }, 
                    {
                        "category_value": "internal", 
                        "category_name": "scope"
                    }
                ]
            }
        }
    ], 
    "provenance_domain": {
        "license": "https://creativecommons.org/licenses/by/4.0/", 
        "name": "Genomics England- Congenital Disorders of Glycosylation", 
        "contributors": [
            {
                "affiliation": "The George Washington University", 
                "contribution": [
                    "createdBy"
                ], 
                "name": "Rahi Navelkar", 
                "email": "rsn13@gwu.edu"
            }, 
            {
                "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"
            }
        ], 
        "created": "2019-11-14T12:15:32.954297", 
        "modified": "2019-12-02T16:44:17.799023", 
        "version": "1.8.24"
    }, 
    "description_domain": {
        "keywords": [
            "Protein", 
            "UniProtKB Accession", 
            "Genomics England", 
            "Panel App"
        ], 
        "pipeline_steps": [
            {
                "name": "make-protein-dataset.py", 
                "step_number": 1, 
                "version": "", 
                "input_list": [
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/source/congenital_disorders_of_glycosylation.tsv", 
                        "filename": "congenital_disorders_of_glycosylation.tsv"
                    }, 
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/reviwed/protein_domap.csv", 
                        "filename": "protein_domap.csv"
                    }
                ], 
                "output_list": [
                    {
                        "uri": "https://data.glygen.org/ln2wwwdata/reviewed/human_protein_genomics_england_disease.csv", 
                        "filename": "human_protein_genomics_england_disease.csv"
                    }
                ], 
                "prerequisite": [], 
                "description": "The input file was downloaded from the Genomics England (https://panelapp.genomicsengland.co.uk/panels/25/) from panel \"Congenital disorders of glycosylation\". A python script was used to process the data. 1. Map the HGNC/gene symbol to UniProtKB canonical accessions from the GlyGen protein masterlist. 2. Only those gene-disease associations are considered where one of the source is \"Expert Review Green\". 3. The MIM ID (Mendelian Inheritance ID) from column \"Phenotypes\" is mapped to DOID (Disease Ontology ID) using input file 2."
            }
        ]
    }, 
    "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/GW-HIVE/glygen-backend-integration/blob/master/pipeline/integrator/make-protein-dataset.py", 
                    "filename": "/make-protein-dataset.py"
                }
            }
        ]
    }, 
    "parametric_domain": [
        {
            "step": "", 
            "value": "", 
            "param": ""
        }
    ], 
    "usability_domain": [
        "The dataset provides Genes involved in the Congenital disorders of glycosylation from Genomics England PanelApp ( PanelApp crowdsources expert knowledge to establish consensus diagnostic gene panels. Antonio Rueda Martin and Eleanor Williams, Rebecca E. Foulger, Sarah Leigh, Louise C. Daugherty, Olivia Niblock, Ivone U. S. Leong, Katherine R. Smith, Oleg Gerasimenko, Eik Haraldsdottir, Ellen Thomas, Richard H. Scott, Emma Baple, Arianna Tucci, Helen Brittain, Anna de Burca, Kristina Iba\u00f1ez, Dalia Kasperaviciute, Damian Smedley, Mark Caulfield, Augusto Rendon & Ellen M. McDonagh. Nat Genet (2019) doi:10.1038/s41588-019-0528-2)"
    ], 
    "etag": "fce22d179c2f7b1f22b98e5d51f552f4", 
    "_id": {
        "$oid": "5dcd8bb4f5c47b0fd850ee6b"
    }
}