{
    "bcoid": "GLY_000235",
    "title": "Mouse Protein Citations (RefSeq)",
    "description": "The Mouse Protein Citations (RefSeq) datasets contains Mus musculus [taxid:10090] UniProtKB accessions ...",
    "filename": "mouse_protein_citations_refseq.csv",
    "filetype": "csv",
    "categories": {
        "species": [
            "Mus musculus"
        ],
        "molecule": [
            "Protein"
        ],
        "tag": [
            "Protein Citations"
        ],
        "file_type": [
            "csv"
        ],
        "status": [
            "Reviewed"
        ],
        "scope": [
            "internal"
        ]
    },
    "minitable": {
        "content": [
            [
                "Q9CQN6-1...",
                "Discovery of..."
            ],
            [
                "P29477-1...",
                "Inducible ni..."
            ]
        ],
        "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"
                }
            ],
            [
                "Q9CQN6-1",
                "Discovery of genes essential for heme biosynthesis through large-scale gene expression analysis.",
                "Cell metabolism",
                "2009 Aug",
                "Nilsson R, Schultz IJ, Pierce EL, Soltis KA, Naranuntarat A, Ward DM, Baughman JM, Paradkar PN, Kingsley PD, Culotta VC, Kaplan J, Palis J, Paw BH, Mootha VK",
                "protein_xref_pubmed",
                "19656490",
                "protein_xref_refseq",
                "NP_079663"
            ],
            [
                "P29477-1",
                "Inducible nitric oxide synthase accelerates nonalcoholic fatty liver disease progression by regulating macrophage autophagy.",
                "Immunity, inflammation and disease",
                "2023-12",
                "Jin Guiyuan, Yao Xiaoying, Liu Dong, Zhang Juan, Zhang Xiaobei, Yang Yonghong, Bi Yanzhen, Zhang Hui, Dong Guanjun, Tang Huixin, Cheng Shumin, Hong Feng, Si Meng",
                "protein_xref_pubmed",
                "38156397",
                "protein_xref_refseq",
                "NP_001300850"
            ]
        ]
    },
    "readme": "Name\n   - Mouse Protein Citations (RefSeq)\n\nVersion\n   - v-2.5.1\n\nCreated on\n  - 2019-09-06T14:06:38.962804\n\nModified on\n   - 2024-05-17T11:23:53.104Z\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 Mouse Protein Citations (RefSeq) datasets contains Mus musculus [taxid:10090] 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: mouse_protein_blacklisted_pmids_uniprotkb.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.5.1/compiled/mouse_protein_blacklisted_pmids_uniprotkb.csv\n      Name: mouse_protein_function_refseq.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/mouse_protein_function_refseq.csv\n   Output Subdomain\n      Name: mouse_protein_citations_refseq.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/mouse_protein_citations_refseq.csv\n      Name: mouse_protein_citations_refseq.stat.csv\n      URI: https:/data.glygen.org/ln2releases/data/v-2.5.1/reviewed/mouse_protein_citations_refseq.stat.csv\n\nFields\n      Name: \"uniprotkb_canonical_ac\"\n     Description: \"\"\n      Unique Values: \"12019\"\n      Name: \"title\"\n     Description: \"Title of the published article in the journal\"\n      Unique Values: \"31349\"\n      Name: \"journal_name\"\n     Description: \"The name of the journal in which the article is published\"\n      Unique Values: \"1467\"\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: \"5323\"\n      Name: \"authors\"\n     Description: \"NA\"\n      Unique Values: \"31269\"\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: \"31349\"\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: \"12019\"",
    "downloadurl": "/ln2data/releases/data/v-2.5.1/reviewed/mouse_protein_citations_refseq.csv"
}