{
    "bcoid": "GLY_000256",
    "title": "Human Protein Citations (RefSeq)",
    "description": "The Human Protein Citations (RefSeq) datasets contains Homo sapiens [taxid:9606] UniProtKB accessions ...",
    "filename_list": "human_protein_citations_refseq.csv",
    "filename": "human_protein_citations_refseq.csv",
    "filetype": "csv",
    "categories": {
        "species": [
            "Homo sapiens"
        ],
        "molecule": [
            "Protein"
        ],
        "tag": [
            "Protein Citations"
        ],
        "file_type": [
            "csv"
        ],
        "status": [
            "Reviewed"
        ],
        "scope": [
            "internal"
        ]
    },
    "minitable": {
        "content": [
            [
                "P0C7P3-1...",
                "SLFN14 mutat..."
            ],
            [
                "Q9BVV6-1...",
                "The CP110-in..."
            ]
        ],
        "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"
                }
            ],
            [
                "P0C7P3-1",
                "SLFN14 mutations underlie thrombocytopenia with excessive bleeding and platelet secretion defects.",
                "The Journal of clinical investigation",
                "2015 Sep",
                "Fletcher SJ, Johnson B, Lowe GC, Bem D, Drake S, Lordkipanidze M, Guiu IS, Dawood B, Rivera J, Simpson MA, Daly ME, Motwani J, Collins PW, Watson SP, Morgan NV",
                "protein_xref_pubmed",
                "26280575",
                "protein_xref_refseq",
                "NP_001123292"
            ],
            [
                "Q9BVV6-1",
                "The CP110-interacting proteins Talpid3 and Cep290 play overlapping and distinct roles in cilia assembly.",
                "The Journal of cell biology",
                "2014 Jan 20",
                "Kobayashi T, Kim S, Lin YC, Inoue T, Dynlacht BD",
                "protein_xref_pubmed",
                "24421332",
                "protein_xref_refseq",
                "NP_001231118"
            ]
        ]
    },
    "readme": "Name\n   - Human Protein Citations (RefSeq)\n\nVersion\n   - v-2.8.1\n\nCreated on\n  - 2019-09-06T14:06:38.983152\n\nModified on\n   - 2024-05-17T11:23:53.899Z\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 Human Protein Citations (RefSeq) datasets contains Homo sapiens [taxid:9606] 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: human_protein_blacklisted_pmids_uniprotkb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.8.1/compiled/human_protein_blacklisted_pmids_uniprotkb.csv\n      Name: human_protein_function_refseq.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.8.1/reviewed/human_protein_function_refseq.csv\n   Output Subdomain\n      Name: human_protein_citations_refseq.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.8.1/reviewed/human_protein_citations_refseq.csv\n      Name: human_protein_citations_refseq.stat.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.8.1/reviewed/human_protein_citations_refseq.stat.csv\n\nFields\n      Name: \"uniprotkb_canonical_ac\"\n     Description: \"\"\n      Unique Values: \"16118\"\n      Name: \"title\"\n     Description: \"Title of the published article in the journal\"\n      Unique Values: \"57821\"\n      Name: \"journal_name\"\n     Description: \"The name of the journal in which the article is published\"\n      Unique Values: \"2557\"\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: \"6421\"\n      Name: \"authors\"\n     Description: \"NA\"\n      Unique Values: \"57623\"\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: \"57824\"\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: \"16118\"",
    "downloadurl": "/ln2data/releases/data/v-2.8.1/reviewed/human_protein_citations_refseq.csv"
}