{
    "bcoid": "GLY_001056",
    "title": "Arabidopsis Protein Citations (UniProtKB)",
    "description": "The Arabidopsis Protein Citations dataset contains publication information for Arabidopsis thaliana [taxid:3702] ...",
    "filename_list": "arabidopsis_protein_citations_uniprotkb.csv",
    "filename": "arabidopsis_protein_citations_uniprotkb.csv",
    "filetype": "csv",
    "categories": {
        "species": [
            "Arabidopsis thaliana"
        ],
        "molecule": [
            "Protein"
        ],
        "tag": [
            "Protein Citations"
        ],
        "file_type": [
            "csv"
        ],
        "status": [
            "Reviewed"
        ],
        "scope": [
            "internal"
        ],
        "file_category": [
            "citations"
        ]
    },
    "minitable": {
        "content": [
            [
                "O64743-1...",
                "Oxidation of..."
            ],
            [
                "O64743-1...",
                "Genetic and ..."
            ]
        ],
        "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"
                }
            ],
            [
                "O64743-1",
                "Oxidation of Monolignols by Members of the Berberine Bridge Enzyme Family Suggests a Role in Plant Cell Wall Metabolism.",
                "The Journal of biological chemistry",
                "2015-07-24",
                "Daniel Bastian, Pavkov-Keller Tea, Steiner Barbara, Dordic Andela, Gutmann Alexander, Nidetzky Bernd, Sensen Christoph W, van der Graaff Eric, Wallner Silvia, Gruber Karl, Macheroux Peter",
                "protein_xref_pubmed",
                "26037923",
                "protein_xref_uniprotkb",
                "O64743"
            ],
            [
                "O64743-1",
                "Genetic and molecular identification of genes required for female gametophyte development and function in Arabidopsis.",
                "Development (Cambridge, England)",
                "2005-02",
                "Pagnussat Gabriela C, Yu Hee-Ju, Ngo Quy A, Rajani Sarojam, Mayalagu Sevugan, Johnson Cameron S, Capron Arnaud, Xie Li-Fen, Ye De, Sundaresan Venkatesan",
                "protein_xref_pubmed",
                "15634699",
                "protein_xref_uniprotkb",
                "O64743"
            ]
        ]
    },
    "readme": "Name\n   - Arabidopsis Protein Citations (UniProtKB)\n\nVersion\n   - v-2.12.1\n\nCreated on\n  - 2024-07-08T14:35:03\n\nModified on\n   - 2024-07-08T14:39:43.781Z\n\nCreated by\n   - Jeet Vora [jeetvora@gwu.edu], Robel Kahsay [rykahsay@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], Jingyue Wu [jingyue.wu@email.gwu.edu], Kate Warner [k.warner1@gwu.edu]\n\nAuthors\n   - Maria Martin [martin@ebi.ac.uk], Preethi Vasudev [pvasudev@ebi.ac.uk]\n\nChecksum\n   - \n\nUsability Domain\n   - The Arabidopsis Protein Citations dataset contains publication information for Arabidopsis thaliana [taxid:3702] UniProtKB accessions that includes PMID, journal name, date and author and title. Publications of large scale studies, and high frequency publications pertaining to genome sequencing, cDNA, chromosomes have been filtered out. If you use this dataset please provide proper attribution to EMBL-EBI-UniProtKB and GlyGen.\n\nDescription Domain\n   Keywords\n   - publications\n   - PMID\n   - journal\n   - citations\n   - authors\n   - UniProtKB\n   - PubMed\n\n   Pipeline Steps\n   - Step-1: The script loads the UniProtKB publication data from the triplestore containing triples from nt file to retrieve the publication information associated with the UniProtKB accession and maps it to the UniProtKB canonical accessions. The script also filters out publications of large scale studies, and high frequency publications pertaining to DNA sequencing, cDNA, chromosomes using blacklisted PMIDs file.\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: all_blacklisted_pmids.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.12.1/compiled/all_blacklisted_pmids.csv\n      Name: arabidopsis_protein_function_uniprotkb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/arabidopsis_protein_function_uniprotkb.csv\n      Name: arabidopsis_protein_ptm_annotation_uniprotkb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/arabidopsis_protein_ptm_annotation_uniprotkb.csv\n      Name: arabidopsis_protein_site_annotation_uniprotkb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/arabidopsis_protein_site_annotation_uniprotkb.csv\n      Name: uniprot-proteome-arabidopsis-thaliana.nt\n      URI: https:/data.glygen.org/ln2downloads/ebi/2026_06_13/uniprot-proteome-arabidopsis-thaliana.nt\n   Output Subdomain\n      Name: arabidopsis_protein_citations_uniprotkb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/arabidopsis_protein_citations_uniprotkb.csv\n      Name: arabidopsis_protein_citations_uniprotkb.stat.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.12.1/reviewed/arabidopsis_protein_citations_uniprotkb.stat.csv\n\nFields\n      Name: \"uniprotkb_canonical_ac\"\"\n     Description: \"\"\n      Unique Values: \"25054\"",
    "downloadurl": "/ln2data/releases/data/v-2.12.1/reviewed/arabidopsis_protein_citations_uniprotkb.csv"
}