{
    "bcoid": "GLY_001440",
    "title": "Arabidopsis Protein Citations (IEDB)",
    "description": "The Arabidopsis Protein Citations (IEDB) dataset contains Arabidopsis thaliana [taxid:3702] UniProtKB ...",
    "filename_list": "arabidopsis_protein_citations_iedb.csv",
    "filename": "arabidopsis_protein_citations_iedb.csv",
    "filetype": "csv",
    "categories": {
        "species": [
            "Arabidopsis thaliana"
        ],
        "molecule": [
            "Protein"
        ],
        "tag": [
            "Protein Citations"
        ],
        "file_type": [
            "csv"
        ],
        "status": [
            "Reviewed"
        ],
        "scope": [
            "internal"
        ]
    },
    "minitable": {
        "content": [
            [
                "Q9LFN8-1...",
                "Endoplasmic ..."
            ],
            [
                "Q2V3L5-1...",
                "Virus-specif..."
            ]
        ],
        "headers": [
            "uniprotkb_ca...",
            "title..."
        ],
        "colwidth": [
            "40%",
            "20%"
        ]
    },
    "sampledata": {
        "type": "table",
        "data": [
            [
                {
                    "type": "string",
                    "label": "uniprotkb_canonical_ac"
                },
                {
                    "type": "string",
                    "label": "title"
                },
                {
                    "type": "string",
                    "label": "journal_name"
                },
                {
                    "type": "string",
                    "label": "publication_date"
                },
                {
                    "type": "string",
                    "label": "authors"
                },
                {
                    "type": "string",
                    "label": "xref_key"
                },
                {
                    "type": "string",
                    "label": "xref_id"
                },
                {
                    "type": "string",
                    "label": "src_xref_key"
                },
                {
                    "type": "string",
                    "label": "src_xref_id"
                }
            ],
            [
                "Q9LFN8-1",
                "Endoplasmic reticulum aminopeptidase associated with antigen processing defines the composition and structure of MHC class I peptide repertoire in normal and virus-infected cells.",
                "Journal of immunology (Baltimore, Md. : 1950)",
                "2010-03-15",
                "Blanchard Nicolas, Kanaseki Takayuki, Escobar Hernando, Delebecque Fr\u00e9d\u00e9ric, Nagarajan Niranjana A, Reyes-Vargas Eduardo, Crockett David K, Raulet David H, Delgado Julio C, Shastri Nilabh",
                "protein_xref_pubmed",
                "20173027",
                "protein_xref_iedb",
                "195941"
            ],
            [
                "Q2V3L5-1",
                "Virus-specific CD4(+) memory-phenotype T cells are abundant in unexposed adults.",
                "Immunity",
                "2013-02-21",
                "Su Laura F, Kidd Brian A, Han Arnold, Kotzin Jonathan J, Davis Mark M",
                "protein_xref_pubmed",
                "23395677",
                "protein_xref_iedb",
                "225961"
            ]
        ]
    },
    "readme": "Name\n   - Arabidopsis Protein Citations (IEDB)\n\nVersion\n   - v-2.9.1\n\nCreated on\n  - \n\nModified on\n   - \n\nCreated by\n   - Jeet Vora [jeetvora@gwu.edu], Karina Martinez [karinamartinez@gwu.edu], Luke Johnson [luke.johnson@email.gwu.edu], Urnisha Bhuiyan [sbhuiyan28@email.gwu.edu], Cyrus Au-Yeung [cyruschauyeung@gwu.edu], Kate Warner [k.warner1@gwu.edu], Mariia Kim [maria.kim@gwu.edu]\n\nAuthors\n   - \n\nChecksum\n   - \n\nUsability Domain\n   - The Arabidopsis Protein Citations (IEDB) dataset contains Arabidopsis thaliana [taxid:3702] UniProtKB accessions mapped to publication information (PMID, title, journal name, publication, date, authors, and the corresponding IEDB accession). If you use this dataset provide proper attribution to IEDB, GlyGen and EBI-EMBL-UniProtKB.\n   - The Immune Epitope Database (IEDB) catalogs experimental data on antibody and T cell epitopes in the context of infectious disease, allergy, autoimmunity and transplantation.\n\nDescription Domain\n   Keywords\n   - PMID\n   - journal\n   - citations\n   - authors\n   - PubMed\n   - IEDB\n\n   Pipeline Steps\n   - Step-1: The script parses the input file for PMIDs and accesses the NCBI Entrez infrastructure to download the publication information of the corresponding PMIDs.\n\nExecution Domain\n   Scripts   - https://github.com/glygener/glygen-backend-integration/blob/master/dataset-maker/make-protein-dataset.py\n\n\n   Software Prerequisites\n      Name: Python\n      Version: 2.7.5\n      URI: https://www.python.org/download/releases/2.7.5/\n\nI/O Domain\n   Input Subdomain\n      Name: arabidopsis_protein_iedb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/arabidopsis_protein_iedb.csv\n   Output Subdomain\n      Name: arabidopsis_protein_citations_iedb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/arabidopsis_protein_citations_iedb.csv\n      Name: arabidopsis_protein_citations_iedb.stat.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/arabidopsis_protein_citations_iedb.stat.csv\n\nFields\n      Name: \"uniprotkb_canonical_ac\"\n     Description: \"\"\n      Unique Values: \"6\"\n      Name: \"title\"\n     Description: \"Title of the published article in the journal\"\n      Unique Values: \"3\"\n      Name: \"journal_name\"\n     Description: \"The name of the journal in which the article is published\"\n      Unique Values: \"3\"\n      Name: \"publication_date\"\n     Description: \"Contains the full date on which the issue of the journal was published. Sourced from MEDLINE/PubMed data element(fields) description.\"\n      Unique Values: \"3\"\n      Name: \"authors\"\n     Description: \"NA\"\n      Unique Values: \"3\"\n      Name: \"xref_key\"\n     Description: \"Source key associated to the  xref_id\"\n      Unique Values: \"1\"\n      Name: \"xref_id\"\n     Description: \"ID of the X-ref\"\n      Unique Values: \"3\"\n      Name: \"src_xref_key\"\n     Description: \"Source Xref  key\"\n      Unique Values: \"1\"\n      Name: \"src_xref_id\"\n     Description: \"Source X-ref ID\"\n      Unique Values: \"6\"",
    "downloadurl": "/ln2data/releases/data/v-2.9.1/reviewed/arabidopsis_protein_citations_iedb.csv"
}