{
    "bcoid": "GLY_001270",
    "title": "Zebrafish Protein Citations (RefSeq)",
    "description": "The Zebrafish Protein Citations (RefSeq) datasets contains Danio rerio [taxid:7955] UniProtKB accessions ...",
    "filename_list": "zebrafish_protein_citations_refseq.csv",
    "filename": "zebrafish_protein_citations_refseq.csv",
    "filetype": "csv",
    "categories": {
        "species": [
            "Danio rerio"
        ],
        "molecule": [
            "Protein"
        ],
        "tag": [
            "Protein Citations"
        ],
        "file_type": [
            "csv"
        ],
        "status": [
            "Reviewed"
        ],
        "scope": [
            "internal"
        ]
    },
    "minitable": {
        "content": [
            [
                "A0A8M1PAA2-1...",
                "Zebrafish sc..."
            ],
            [
                "Q5I2A9-1...",
                "Control of e..."
            ]
        ],
        "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"
                }
            ],
            [
                "A0A8M1PAA2-1",
                "Zebrafish scube1 (signal peptide-CUB (complement protein C1r/C1s, Uegf, and Bmp1)-EGF (epidermal growth factor) domain-containing protein 1) is involved in primitive hematopoiesis.",
                "The Journal of biological chemistry",
                "2013-02-15",
                "Tsao Ku-Chi, Tu Cheng-Fen, Lee Shyh-Jye, Yang Ruey-Bing",
                "protein_xref_pubmed",
                "23271740",
                "protein_xref_refseq",
                "NP_001315528"
            ],
            [
                "Q5I2A9-1",
                "Control of extracellular matrix assembly along tissue boundaries via Integrin and Eph/Ephrin signaling.",
                "Development (Cambridge, England)",
                "2009-09",
                "J\u00fclich D\u00f6rthe, Mould A Paul, Koper Ewa, Holley Scott A",
                "protein_xref_pubmed",
                "19641014",
                "protein_xref_refseq",
                "NP_001004288"
            ]
        ]
    },
    "readme": "Name\n   - Zebrafish Protein Citations (RefSeq)\n\nVersion\n   - v-2.9.1\n\nCreated on\n  - \n\nModified on\n   - \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], Kate Warner [k.warner1@gwu.edu]\n\nAuthors\n   - Not Applicable []\n\nChecksum\n   - \n\nUsability Domain\n   - The Zebrafish Protein Citations (RefSeq) datasets contains Danio rerio [taxid:7955] UniProtKB accessions mapped to publication information (PMID, title, journal name, publication, date, authors of corresponding RefSeq accessions). If you use this dataset please provide proper attribution to NCBI RefSeq and GlyGen. \n\nDescription Domain\n   Keywords\n   - PMID\n   - journal\n   - citations\n   - authors\n   - RefSeq\n   - PubMed\n\n   Pipeline Steps\n   - Step-1: The script parses the NCBI RefSeq gpff files to extract the publication information. The script also maps RefSeq accessions to the UniProtKB Canonical accessions. \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-2.9.1/compiled/all_blacklisted_pmids.csv\n      Name: zebrafish_protein_function_refseq.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/zebrafish_protein_function_refseq.csv\n   Output Subdomain\n      Name: zebrafish_protein_citations_refseq.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/zebrafish_protein_citations_refseq.csv\n      Name: zebrafish_protein_citations_refseq.stat.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.9.1/reviewed/zebrafish_protein_citations_refseq.stat.csv\n\nFields\n      Name: \"uniprotkb_canonical_ac\"\n     Description: \"\"\n      Unique Values: \"2491\"\n      Name: \"title\"\n     Description: \"Title of the published article in the journal\"\n      Unique Values: \"3019\"\n      Name: \"journal_name\"\n     Description: \"The name of the journal in which the article is published\"\n      Unique Values: \"404\"\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: \"1657\"\n      Name: \"authors\"\n     Description: \"NA\"\n      Unique Values: \"3005\"\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: \"3019\"\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: \"2491\"",
    "downloadurl": "/ln2data/releases/data/v-2.9.1/reviewed/zebrafish_protein_citations_refseq.csv"
}