{
    "bcoid": "GLY_000400",
    "title": "Rat Protein Sequence Info (UniProtKB)",
    "description": "The Rat Protein Sequence Info (UniProtKB) dataset contains rat [taxid:10116] UniProtKB protein fasta ...",
    "filename": "rat_protein_sequenceinfo.csv",
    "filetype": "csv",
    "categories": {
        "species": "Rattus norvegicus",
        "molecule": "Protein",
        "tags": "All Protein Sequence Info",
        "file_type": "csv",
        "status": "Reviewed",
        "scope": "internal"
    },
    "minitable": {
        "content": [
            [
                "D3ZB44-1...",
                "D3ZB44-1..."
            ],
            [
                "D3ZXJ3-1...",
                "D3ZB41-1..."
            ]
        ],
        "headers": [
            "uniprotkb_ca...",
            "uniprotkb_is..."
        ],
        "colwidth": [
            "40%",
            "20%"
        ]
    },
    "sampledata": {
        "type": "table",
        "data": [
            [
                {
                    "type": "string",
                    "label": "uniprotkb_canonical_ac"
                },
                {
                    "type": "string",
                    "label": "uniprotkb_isoform_ac"
                },
                {
                    "type": "string",
                    "label": "protein_existence"
                },
                {
                    "type": "string",
                    "label": "sequence_version"
                },
                {
                    "type": "string",
                    "label": "sequence_header"
                }
            ],
            [
                "D3ZB44-1",
                "D3ZB44-1",
                "4",
                "1",
                "tr|D3ZB44-1|D3ZB44_RAT Isoform 1 of Zinc finger protein 324  OS=Rattus norvegicus OX=10116 GN=Zfp324 PE=4 SV=1 CANONICAL=D3ZB44-1"
            ],
            [
                "D3ZXJ3-1",
                "D3ZB41-1",
                "4",
                "2",
                "tr|D3ZB41-1|D3ZB41_RAT Computationally mapped isoform of Zinc finger protein 78  OS=Rattus norvegicus OX=10116 GN=Zfp78 PE=4 SV=2 CANONICAL=D3ZXJ3-1"
            ],
            [
                "Q64627-1",
                "Q5BK97-1",
                "1",
                "1",
                "tr|Q5BK97-1|NEUR2_RAT Computationally mapped isoform of Sialidase-2  OS=Rattus norvegicus OX=10116 GN=Neu2 PE=1 SV=1 CANONICAL=Q64627-1"
            ],
            [
                "Q64627-1",
                "Q64627-1",
                "1",
                "1",
                "sp|Q64627-1|NEUR2_RAT Isoform 1 of Sialidase-2  OS=Rattus norvegicus OX=10116 GN=Neu2 PE=1 SV=1 CANONICAL=Q64627-1"
            ],
            [
                "D3ZB48-1",
                "D3ZB48-1",
                "1",
                "1",
                "tr|D3ZB48-1|D3ZB48_RAT Isoform 1 of Hook microtubule-tethering protein 1  OS=Rattus norvegicus OX=10116 GN=Hook1 PE=1 SV=1 CANONICAL=D3ZB48-1"
            ],
            [
                "Q5RJZ1-1",
                "A0A0G2JVG0-1",
                "3",
                "1",
                "tr|A0A0G2JVG0-1|A0A0G2JVG0_RAT Computationally mapped isoform of Regulator of telomere elongation helicase 1  OS=Rattus norvegicus OX=10116 GN=Rtel1 PE=3 SV=1 CANONICAL=Q5RJZ1-1"
            ],
            [
                "Q04462-1",
                "A0A0G2JVG2-1",
                "1",
                "1",
                "tr|A0A0G2JVG2-1|A0A0G2JVG2_RAT Computationally mapped isoform of Valine--tRNA ligase  OS=Rattus norvegicus OX=10116 GN=Vars PE=1 SV=1 CANONICAL=Q04462-1"
            ],
            [
                "P11980-1",
                "A0A0G2JVG3-1",
                "1",
                "1",
                "tr|A0A0G2JVG3-1|A0A0G2JVG3_RAT Computationally mapped isoform of Pyruvate kinase  OS=Rattus norvegicus OX=10116 GN=Pkm PE=1 SV=1 CANONICAL=P11980-1"
            ],
            [
                "Q9WVK3-1",
                "A0A0G2JVG4-1",
                "1",
                "1",
                "tr|A0A0G2JVG4-1|A0A0G2JVG4_RAT Computationally mapped isoform of Peroxisomal trans-2-enoyl-CoA reductase  OS=Rattus norvegicus OX=10116 GN=Pecr PE=1 SV=1 CANONICAL=Q9WVK3-1"
            ],
            [
                "A0A0G2JVG5-1",
                "A0A0G2JVG5-1",
                "4",
                "1",
                "tr|A0A0G2JVG5-1|A0A0G2JVG5_RAT Isoform 1 of Trichohyalin-like  OS=Rattus norvegicus OX=10116 GN=LOC102552378 PE=4 SV=1 CANONICAL=A0A0G2JVG5-1"
            ],
            [
                "A0A0G2JVG6-1",
                "A0A0G2JVG6-1",
                "4",
                "1",
                "tr|A0A0G2JVG6-1|A0A0G2JVG6_RAT Isoform 1 of Uncharacterized protein  OS=Rattus norvegicus OX=10116 GN= PE=4 SV=1 CANONICAL=A0A0G2JVG6-1"
            ],
            [
                "P30938-1",
                "A0A0G2K3H3-1",
                "1",
                "1",
                "tr|A0A0G2K3H3-1|SSR5_RAT Computationally mapped isoform of Somatostatin receptor type 5  OS=Rattus norvegicus OX=10116 GN=Sstr5 PE=1 SV=1 CANONICAL=P30938-1"
            ],
            [
                "P30938-1",
                "G3V8G3-1",
                "1",
                "1",
                "tr|G3V8G3-1|SSR5_RAT Computationally mapped isoform of Somatostatin receptor type 5  OS=Rattus norvegicus OX=10116 GN=Sstr5 PE=1 SV=1 CANONICAL=P30938-1"
            ],
            [
                "P30938-1",
                "P30938-1",
                "1",
                "2",
                "sp|P30938-1|SSR5_RAT Isoform 1 of Somatostatin receptor type 5  OS=Rattus norvegicus OX=10116 GN=Sstr5 PE=1 SV=2 CANONICAL=P30938-1"
            ],
            [
                "A0A0G2JVG8-1",
                "A0A0G2JVG8-1",
                "3",
                "1",
                "tr|A0A0G2JVG8-1|A0A0G2JVG8_RAT Isoform 1 of Olfactory receptor  OS=Rattus norvegicus OX=10116 GN=Olr410 PE=3 SV=1 CANONICAL=A0A0G2JVG8-1"
            ],
            [
                "Q07969-1",
                "A0A096MIY7-1",
                "1",
                "1",
                "tr|A0A096MIY7-1|CD36_RAT Computationally mapped isoform of Platelet glycoprotein 4  OS=Rattus norvegicus OX=10116 GN=Cd36 PE=1 SV=1 CANONICAL=Q07969-1"
            ],
            [
                "Q07969-1",
                "A0A096MJ39-1",
                "1",
                "1",
                "tr|A0A096MJ39-1|CD36_RAT Computationally mapped isoform of Platelet glycoprotein 4  OS=Rattus norvegicus OX=10116 GN=Cd36 PE=1 SV=1 CANONICAL=Q07969-1"
            ],
            [
                "Q07969-1",
                "A0A0G2JYC5-1",
                "1",
                "1",
                "tr|A0A0G2JYC5-1|CD36_RAT Computationally mapped isoform of Platelet glycoprotein 4  OS=Rattus norvegicus OX=10116 GN=Cd36 PE=1 SV=1 CANONICAL=Q07969-1"
            ],
            [
                "Q07969-1",
                "D4A8Z2-1",
                "1",
                "3",
                "tr|D4A8Z2-1|CD36_RAT Computationally mapped isoform of Platelet glycoprotein 4  OS=Rattus norvegicus OX=10116 GN=Cd36 PE=1 SV=3 CANONICAL=Q07969-1"
            ],
            [
                "Q07969-1",
                "Q07969-1",
                "1",
                "3",
                "sp|Q07969-1|CD36_RAT Isoform 1 of Platelet glycoprotein 4  OS=Rattus norvegicus OX=10116 GN=Cd36 PE=1 SV=3 CANONICAL=Q07969-1"
            ],
            [
                "Q07969-1",
                "Q6IMX5-1",
                "1",
                "1",
                "tr|Q6IMX5-1|CD36_RAT Computationally mapped isoform of Platelet glycoprotein 4  OS=Rattus norvegicus OX=10116 GN=Cd36 PE=1 SV=1 CANONICAL=Q07969-1"
            ],
            [
                "A0A096MJY8-1",
                "A0A096MJY8-1",
                "3",
                "1",
                "tr|A0A096MJY8-1|A0A096MJY8_RAT Isoform 1 of Acetyl-CoA acetyltransferase 2-like 1  OS=Rattus norvegicus OX=10116 GN=Acat2l1 PE=3 SV=1 CANONICAL=A0A096MJY8-1"
            ],
            [
                "A0A096MJY8-1",
                "D3ZKB6-1",
                "3",
                "3",
                "tr|D3ZKB6-1|A0A096MJY8_RAT Computationally mapped isoform of Acetyl-CoA acetyltransferase 2-like 1  OS=Rattus norvegicus OX=10116 GN=Acat2l1 PE=3 SV=3 CANONICAL=A0A096MJY8-1"
            ],
            [
                "Q9QYG8-1",
                "A0A096MJY5-1",
                "3",
                "1",
                "tr|A0A096MJY5-1|A0A096MJY5_RAT Computationally mapped isoform of Uridine/cytidine kinase  OS=Rattus norvegicus OX=10116 GN=Uck2 PE=3 SV=1 CANONICAL=Q9QYG8-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MJ09-1",
                "1",
                "1",
                "tr|A0A096MJ09-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MJ23-1",
                "1",
                "1",
                "tr|A0A096MJ23-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MJD9-1",
                "1",
                "1",
                "tr|A0A096MJD9-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MJY4-1",
                "1",
                "1",
                "sp|A0A096MJY4-1|MEF2C_RAT Isoform 1 of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MJZ6-1",
                "1",
                "1",
                "tr|A0A096MJZ6-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MK01-1",
                "1",
                "1",
                "tr|A0A096MK01-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MK43-1",
                "1",
                "1",
                "tr|A0A096MK43-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MK88-1",
                "1",
                "1",
                "tr|A0A096MK88-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "A0A096MJY4-1",
                "A0A096MKI4-1",
                "1",
                "1",
                "tr|A0A096MKI4-1|MEF2C_RAT Computationally mapped isoform of Myocyte-specific enhancer factor 2C  OS=Rattus norvegicus OX=10116 GN=Mef2c PE=1 SV=1 CANONICAL=A0A096MJY4-1"
            ],
            [
                "B2GV58-1",
                "A0A096MJY7-1",
                "4",
                "1",
                "tr|A0A096MJY7-1|A0A096MJY7_RAT Computationally mapped isoform of LSM14A mRNA-processing body assembly factor  OS=Rattus norvegicus OX=10116 GN=Lsm14a PE=4 SV=1 CANONICAL=B2GV58-1"
            ],
            [
                "A0A0G2JTB2-1",
                "A0A096MJY6-1",
                "1",
                "1",
                "tr|A0A096MJY6-1|A0A096MJY6_RAT Computationally mapped isoform of 1,4-alpha-glucan-branching enzyme 1  OS=Rattus norvegicus OX=10116 GN=Gbe1 PE=1 SV=1 CANONICAL=A0A0G2JTB2-1"
            ],
            [
                "A0A096MJY1-1",
                "A0A096MJY1-1",
                "1",
                "1",
                "tr|A0A096MJY1-1|A0A096MJY1_RAT Isoform 1 of Similar to Gpc6 protein (Predicted)  OS=Rattus norvegicus OX=10116 GN=Gpc6 PE=1 SV=1 CANONICAL=A0A096MJY1-1"
            ],
            [
                "A0A0G2KA92-1",
                "A0A096MJY0-1",
                "1",
                "1",
                "tr|A0A096MJY0-1|A0A096MJY0_RAT Computationally mapped isoform of DNA helicase  OS=Rattus norvegicus OX=10116 GN=Chd2 PE=1 SV=1 CANONICAL=A0A0G2KA92-1"
            ],
            [
                "P56558-1",
                "A0A096MJY3-1",
                "1",
                "1",
                "tr|A0A096MJY3-1|A0A096MJY3_RAT Computationally mapped isoform of UDP-N-acetylglucosamine--peptide N-acetylglucosaminyltransferase 110 kDa subunit  OS=Rattus norvegicus OX=10116 GN=Ogt PE=1 SV=1 CANONICAL=P56558-1"
            ],
            [
                "A0A096MJY2-1",
                "A0A096MJY2-1",
                "3",
                "1",
                "tr|A0A096MJY2-1|A0A096MJY2_RAT Isoform 1 of Stomatin-like 3  OS=Rattus norvegicus OX=10116 GN=Stoml3 PE=3 SV=1 CANONICAL=A0A096MJY2-1"
            ],
            [
                "A0A096MJY2-1",
                "D4A100-1",
                "3",
                "2",
                "tr|D4A100-1|A0A096MJY2_RAT Computationally mapped isoform of Stomatin-like 3  OS=Rattus norvegicus OX=10116 GN=Stoml3 PE=3 SV=2 CANONICAL=A0A096MJY2-1"
            ],
            [
                "F7ES73-1",
                "F7ES73-1",
                "1",
                "1",
                "tr|F7ES73-1|F7ES73_RAT Isoform 1 of Negative regulator of ubiquitin-like proteins 1  OS=Rattus norvegicus OX=10116 GN=Nub1 PE=1 SV=1 CANONICAL=F7ES73-1"
            ],
            [
                "P01946-1",
                "P01946-1",
                "1",
                "3",
                "sp|P01946-1|HBA_RAT Isoform 1 of Hemoglobin subunit alpha-1/2  OS=Rattus norvegicus OX=10116 GN=Hba1 PE=1 SV=3 CANONICAL=P01946-1"
            ],
            [
                "Q63358-1",
                "Q4W1H3-1",
                "1",
                "1",
                "tr|Q4W1H3-1|Q4W1H3_RAT Computationally mapped isoform of Unconventional myosin-IXb  OS=Rattus norvegicus OX=10116 GN=Myo9b PE=1 SV=1 CANONICAL=Q63358-1"
            ],
            [
                "P30937-1",
                "P30937-1",
                "1",
                "1",
                "sp|P30937-1|SSR4_RAT Isoform 1 of Somatostatin receptor type 4  OS=Rattus norvegicus OX=10116 GN=Sstr4 PE=1 SV=1 CANONICAL=P30937-1"
            ],
            [
                "O35152-1",
                "A0A0G2JSM4-1",
                "1",
                "1",
                "tr|A0A0G2JSM4-1|BET1L_RAT Computationally mapped isoform of BET1-like protein  OS=Rattus norvegicus OX=10116 GN=Bet1l PE=1 SV=1 CANONICAL=O35152-1"
            ],
            [
                "O35152-1",
                "O35152-1",
                "1",
                "1",
                "sp|O35152-1|BET1L_RAT Isoform 1 of BET1-like protein  OS=Rattus norvegicus OX=10116 GN=Bet1l PE=1 SV=1 CANONICAL=O35152-1"
            ],
            [
                "P30936-1",
                "P30936-1",
                "1",
                "1",
                "sp|P30936-1|SSR3_RAT Isoform 1 of Somatostatin receptor type 3  OS=Rattus norvegicus OX=10116 GN=Sstr3 PE=1 SV=1 CANONICAL=P30936-1"
            ],
            [
                "O35158-1",
                "G3V7K7-1",
                "1",
                "1",
                "tr|G3V7K7-1|CDON_RAT Computationally mapped isoform of Cell adhesion molecule-related/down-regulated by oncogenes  OS=Rattus norvegicus OX=10116 GN=Cdon PE=1 SV=1 CANONICAL=O35158-1"
            ],
            [
                "O35158-1",
                "O35158-1",
                "1",
                "2",
                "sp|O35158-1|CDON_RAT Isoform 1 of Cell adhesion molecule-related/down-regulated by oncogenes  OS=Rattus norvegicus OX=10116 GN=Cdon PE=1 SV=2 CANONICAL=O35158-1"
            ],
            [
                "O35158-1",
                "O35158-2",
                "1",
                "2",
                "sp|O35158-2|CDON_RAT Isoform 2 of Cell adhesion molecule-related/down-regulated by oncogenes  OS=Rattus norvegicus OX=10116 GN=Cdon PE=1 SV=2 CANONICAL=O35158-1"
            ],
            [
                "P02761-1",
                "P02761-1",
                "1",
                "1",
                "sp|P02761-1|MUP_RAT Isoform 1 of Major urinary protein  OS=Rattus norvegicus OX=10116 GN= PE=1 SV=1 CANONICAL=P02761-1"
            ],
            [
                "Q9QYH1-1",
                "D3ZUC7-1",
                "1",
                "3",
                "tr|D3ZUC7-1|MPP4_RAT Computationally mapped isoform of MAGUK p55 subfamily member 4  OS=Rattus norvegicus OX=10116 GN=Mpp4 PE=1 SV=3 CANONICAL=Q9QYH1-1"
            ],
            [
                "Q9QYH1-1",
                "F1MA44-1",
                "1",
                "2",
                "tr|F1MA44-1|MPP4_RAT Computationally mapped isoform of MAGUK p55 subfamily member 4  OS=Rattus norvegicus OX=10116 GN=Mpp4 PE=1 SV=2 CANONICAL=Q9QYH1-1"
            ],
            [
                "Q9QYH1-1",
                "Q9QYH1-1",
                "1",
                "1",
                "sp|Q9QYH1-1|MPP4_RAT Isoform 1 of MAGUK p55 subfamily member 4  OS=Rattus norvegicus OX=10116 GN=Mpp4 PE=1 SV=1 CANONICAL=Q9QYH1-1"
            ],
            [
                "Q9QYH1-1",
                "Q9QYH1-2",
                "1",
                "1",
                "sp|Q9QYH1-2|MPP4_RAT Isoform 2 of MAGUK p55 subfamily member 4  OS=Rattus norvegicus OX=10116 GN=Mpp4 PE=1 SV=1 CANONICAL=Q9QYH1-1"
            ],
            [
                "Q9QYH1-1",
                "Q9QYH1-3",
                "1",
                "1",
                "sp|Q9QYH1-3|MPP4_RAT Isoform 3 of MAGUK p55 subfamily member 4  OS=Rattus norvegicus OX=10116 GN=Mpp4 PE=1 SV=1 CANONICAL=Q9QYH1-1"
            ],
            [
                "P02764-1",
                "A0A0H2UHF8-1",
                "2",
                "1",
                "tr|A0A0H2UHF8-1|A1AG_RAT Computationally mapped isoform of Alpha-1-acid glycoprotein  OS=Rattus norvegicus OX=10116 GN=Orm1 PE=2 SV=1 CANONICAL=P02764-1"
            ],
            [
                "P02764-1",
                "P02764-1",
                "2",
                "1",
                "sp|P02764-1|A1AG_RAT Isoform 1 of Alpha-1-acid glycoprotein  OS=Rattus norvegicus OX=10116 GN=Orm1 PE=2 SV=1 CANONICAL=P02764-1"
            ],
            [
                "P02767-1",
                "P02767-1",
                "1",
                "1",
                "sp|P02767-1|TTHY_RAT Isoform 1 of Transthyretin  OS=Rattus norvegicus OX=10116 GN=Ttr PE=1 SV=1 CANONICAL=P02767-1"
            ],
            [
                "D4A4W8-1",
                "D4A4W8-1",
                "4",
                "1",
                "tr|D4A4W8-1|D4A4W8_RAT Isoform 1 of T, brachyury homolog (Mouse) (Predicted), isoform CRA_b  OS=Rattus norvegicus OX=10116 GN=Tbxt PE=4 SV=1 CANONICAL=D4A4W8-1"
            ],
            [
                "D4A4W9-1",
                "D4A4W9-1",
                "3",
                "1",
                "tr|D4A4W9-1|D4A4W9_RAT Isoform 1 of 60S ribosomal protein L34  OS=Rattus norvegicus OX=10116 GN= PE=3 SV=1 CANONICAL=D4A4W9-1"
            ],
            [
                "D4A4W0-1",
                "D4A4W0-1",
                "3",
                "1",
                "tr|D4A4W0-1|D4A4W0_RAT Isoform 1 of Olfactory receptor  OS=Rattus norvegicus OX=10116 GN=Olr338 PE=3 SV=1 CANONICAL=D4A4W0-1"
            ],
            [
                "D4A4W1-1",
                "A0A096MJT8-1",
                "3",
                "1",
                "tr|A0A096MJT8-1|D4A4W1_RAT Computationally mapped isoform of Family with sequence similarity 174, member B  OS=Rattus norvegicus OX=10116 GN=Fam174b PE=3 SV=1 CANONICAL=D4A4W1-1"
            ],
            [
                "D4A4W1-1",
                "D4A4W1-1",
                "3",
                "2",
                "tr|D4A4W1-1|D4A4W1_RAT Isoform 1 of Family with sequence similarity 174, member B  OS=Rattus norvegicus OX=10116 GN=Fam174b PE=3 SV=2 CANONICAL=D4A4W1-1"
            ],
            [
                "D4A4W2-1",
                "D4A4W2-1",
                "3",
                "3",
                "tr|D4A4W2-1|D4A4W2_RAT Isoform 1 of Multidrug and toxin extrusion protein  OS=Rattus norvegicus OX=10116 GN=Slc47a2 PE=3 SV=3 CANONICAL=D4A4W2-1"
            ],
            [
                "F7EN52-1",
                "F7EN52-1",
                "2",
                "2",
                "tr|F7EN52-1|F7EN52_RAT Isoform 1 of Cyp46a1 protein  OS=Rattus norvegicus OX=10116 GN=Cyp46a1 PE=2 SV=2 CANONICAL=F7EN52-1"
            ],
            [
                "F7EN52-1",
                "A0A0G2JWG7-1",
                "2",
                "1",
                "tr|A0A0G2JWG7-1|F7EN52_RAT Computationally mapped isoform of Cyp46a1 protein  OS=Rattus norvegicus OX=10116 GN=Cyp46a1 PE=2 SV=1 CANONICAL=F7EN52-1"
            ],
            [
                "D4A4W4-1",
                "D4A4W4-1",
                "4",
                "1",
                "tr|D4A4W4-1|D4A4W4_RAT Isoform 1 of Abhydrolase domain containing 9 (Predicted)  OS=Rattus norvegicus OX=10116 GN=Ephx3 PE=4 SV=1 CANONICAL=D4A4W4-1"
            ],
            [
                "D4A4W5-1",
                "D4A4W5-1",
                "3",
                "1",
                "tr|D4A4W5-1|D4A4W5_RAT Isoform 1 of Phospholipid-transporting ATPase  OS=Rattus norvegicus OX=10116 GN=Atp10b PE=3 SV=1 CANONICAL=D4A4W5-1"
            ],
            [
                "D4A4W6-1",
                "D4A4W6-1",
                "1",
                "1",
                "tr|D4A4W6-1|D4A4W6_RAT Isoform 1 of SRA stem-loop-interacting RNA-binding protein  OS=Rattus norvegicus OX=10116 GN=Slirp PE=1 SV=1 CANONICAL=D4A4W6-1"
            ],
            [
                "D4A4W7-1",
                "D4A4W7-1",
                "1",
                "1",
                "tr|D4A4W7-1|D4A4W7_RAT Isoform 1 of Mitochondria-localized glutamic acid-rich protein  OS=Rattus norvegicus OX=10116 GN=Mgarp PE=1 SV=1 CANONICAL=D4A4W7-1"
            ],
            [
                "G3V6I3-1",
                "G3V6I3-1",
                "1",
                "1",
                "tr|G3V6I3-1|G3V6I3_RAT Isoform 1 of Coiled-coil domain containing 49 (Predicted)  OS=Rattus norvegicus OX=10116 GN=Cwc25 PE=1 SV=1 CANONICAL=G3V6I3-1"
            ],
            [
                "P21932-1",
                "G3V6I2-1",
                "3",
                "1",
                "tr|G3V6I2-1|G3V6I2_RAT Computationally mapped isoform of Adenylate cyclase type 3  OS=Rattus norvegicus OX=10116 GN=Adcy3 PE=3 SV=1 CANONICAL=P21932-1"
            ],
            [
                "Q8K4K5-1",
                "G3V6I1-1",
                "1",
                "2",
                "tr|G3V6I1-1|G3V6I1_RAT Computationally mapped isoform of Lethal(2) giant larvae protein homolog 1  OS=Rattus norvegicus OX=10116 GN=Llgl1 PE=1 SV=2 CANONICAL=Q8K4K5-1"
            ],
            [
                "Q4L0E8-1",
                "G3V6I0-1",
                "1",
                "1",
                "tr|G3V6I0-1|G3V6I0_RAT Computationally mapped isoform of Regulator of G-protein signaling 18  OS=Rattus norvegicus OX=10116 GN=Rgs18 PE=1 SV=1 CANONICAL=Q4L0E8-1"
            ],
            [
                "G3V6I6-1",
                "G3V6I6-1",
                "3",
                "1",
                "tr|G3V6I6-1|G3V6I6_RAT Isoform 1 of Adiponectin receptor 1  OS=Rattus norvegicus OX=10116 GN=Adipor1 PE=3 SV=1 CANONICAL=G3V6I6-1"
            ],
            [
                "A0A0G2K5E4-1",
                "G3V6I5-1",
                "1",
                "2",
                "tr|G3V6I5-1|G3V6I5_RAT Computationally mapped isoform of DnaJ heat shock protein family (Hsp40) member A3  OS=Rattus norvegicus OX=10116 GN=Dnaja3 PE=1 SV=2 CANONICAL=A0A0G2K5E4-1"
            ],
            [
                "G3V6I4-1",
                "G3V6I4-1",
                "1",
                "1",
                "tr|G3V6I4-1|G3V6I4_RAT Isoform 1 of Mitochondrial amidoxime reducing component 1  OS=Rattus norvegicus OX=10116 GN=Marc1 PE=1 SV=1 CANONICAL=G3V6I4-1"
            ],
            [
                "P12749-1",
                "G3V6I9-1",
                "1",
                "1",
                "tr|G3V6I9-1|G3V6I9_RAT Computationally mapped isoform of RCG33968, isoform CRA_a  OS=Rattus norvegicus OX=10116 GN=Rpl26 PE=1 SV=1 CANONICAL=P12749-1"
            ],
            [
                "O08812-1",
                "G3V6I8-1",
                "3",
                "1",
                "tr|G3V6I8-1|G3V6I8_RAT Computationally mapped isoform of Cationic amino acid transporter 3  OS=Rattus norvegicus OX=10116 GN=Slc7a3 PE=3 SV=1 CANONICAL=O08812-1"
            ],
            [
                "Q767I8-1",
                "M0RBC1-1",
                "4",
                "1",
                "tr|M0RBC1-1|M0RBC1_RAT Computationally mapped isoform of Protocadherin alpha-4  OS=Rattus norvegicus OX=10116 GN=Pcdha4 PE=4 SV=1 CANONICAL=Q767I8-1"
            ],
            [
                "P39020-1",
                "G3V6T9-1",
                "2",
                "1",
                "tr|G3V6T9-1|MEOX2_RAT Computationally mapped isoform of Homeobox protein MOX-2  OS=Rattus norvegicus OX=10116 GN=Meox2 PE=2 SV=1 CANONICAL=P39020-1"
            ],
            [
                "P39020-1",
                "P39020-1",
                "2",
                "1",
                "sp|P39020-1|MEOX2_RAT Isoform 1 of Homeobox protein MOX-2  OS=Rattus norvegicus OX=10116 GN=Meox2 PE=2 SV=1 CANONICAL=P39020-1"
            ],
            [
                "Q0HA29-1",
                "A0A0U1RRT0-1",
                "2",
                "1",
                "tr|A0A0U1RRT0-1|NMNA2_RAT Computationally mapped isoform of Nicotinamide/nicotinic acid mononucleotide adenylyltransferase 2  OS=Rattus norvegicus OX=10116 GN=Nmnat2 PE=2 SV=1 CANONICAL=Q0HA29-1"
            ]
        ]
    },
    "readme": "Name\n   - Rat Protein Sequence Info (UniProtKB)\n\nVersion\n   - 1.11.2\n\nCreated on\n  - 2020-03-15T14:37:18.669955\n\nModified on\n   - 2020-03-15T15:15:57.622792\n\nCreated by\n   - Jeet Vora [jeetvora@gwu.edu], Rahi Navelkar [rsn13@gwu.edu], Robel Kahsay [rykahsay@gwu.edu], Karina Martinez [karinamartinez@gwu.edu], Amanda Bell [amandab2140@gwu.edu], Jonathon Keeney [keeneyjg@gwu.edu], Millicent Quartey [mquartey@email.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 Rat Protein Sequence Info (UniProtKB) dataset contains rat [taxid:10116] UniProtKB protein fasta sequence information that includes sequence version and fasta header. The dataset is derived from 2019-09 UniProtKB release. If you use this dataset please provide proper attribution to EMBL-EBI-UniProtKB and GlyGen\n\nDescription Domain\n   Keywords\n   - protein sequence info\n   - fasta header\n   - UniProtKB\n\n   Pipeline Steps\n   - Step-1: The script loads the UniProtKB sequence for all rat isoform accessions data from the triplestore containing triples from nt file and extracts it into a dataset in a fasta format\n\nExecution Domain\n   Scripts   - https://github.com/GW-HIVE/glygen-backend-integration/blob/master/pipeline/integrator/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: uniprot-proteome-rattus-norvegicus.nt\n      URI: https://data.glygen.org/ln2wwwdata/source/uniprot-proteome-rattus-norvegicus.nt\n   Output Subdomain\n      Name: rat_protein_sequenceinfo.csv\n      URI: http://data.glygen.org/ln2wwwdata/reviewed/rat_protein_sequenceinfo.csv\n      Name: rat_protein_sequenceinfo.stat.csv\n      URI: http://data.glygen.org/ln2wwwdata/reviewed/rat_protein_rat_sequenceinfo.stat.csv\n\nFields\n      Name: \"uniprotkb_canonical_ac\"\n     Description: \"\"\n      Unique Values: \"16\"\n      Name: \"motif_start\"\n     Description: \"\"\n      Unique Values: \"23\"\n      Name: \"filter_flags\"\n     Description: \"Filter tags/flags assigned to the mutation entries that have passed the mutation filter criteria\"\n      Unique Values: \"19\"\n      Name: \"sift_prediction\"\n     Description: \"The SIFT qualitative prediction is derived from this score such that substitutions with a score < 0.05 are called 'deleterious' and all others are called 'tolerated\"\n      Unique Values: \"3\"\n      Name: \"somatic_status\"\n     Description: \"The status of the variant whether it is germline (somatic status =0) or somatic (somatic status =1) variant\"\n      Unique Values: \"1\"\n      Name: \"chr_pos\"\n     Description: \"Chromosome Position of a gene\"\n      Unique Values: \"19\"\n      Name: \"chr_id\"\n     Description: \"\"\n      Unique Values: \"10\"\n      Name: \"dbsnp_id\"\n     Description: \"\"A reference SNP ID number\"\n      Unique Values: \"19\"\n      Name: \"evidence_ECO:0000313\"\n     Description: \"A type of evidence that is imported information used in an automatic assertion\"\n      Unique Values: \"1\"\n      Name: \"xref_id\"\n     Description: \"ID of the X-ref\"\n      Unique Values: \"19\"\n      Name: \"mim_id\"\n     Description: \"Disease ID from the OMIM database\"\n      Unique Values: \"1\"\n      Name: \"description\"\n     Description: \"\"\n      Unique Values: \"1\"\n      Name: \"begin_aa_pos\"\n     Description: \"Start position of the amino acid in the sequence of a protein in UniProtKB\"\n      Unique Values: \"18\"\n      Name: \"cytogenic_band\"\n     Description: \"One of the subregions of a chromosome visible microscopically after special staining\"\n      Unique Values: \"14\"\n      Name: \"aa_pos\"\n     Description: \"Position of the amino acid in canonical sequence of protein in the UniProtKB database\"\n      Unique Values: \"18\"\n      Name: \"ref_aa\"\n     Description: \"Reference amino acid residue\"\n      Unique Values: \"9\"\n      Name: \"sift_score\"\n     Description: \"SIFT predicts whether an amino acid substitution is likely to affect protein function based on sequence homology and the physico-chemical similarity between the alternate amino acids. The data we provide for each amino acid substitution is a score and a qualitative prediction (either 'tolerated' or 'deleterious'). The score is the normalized probability that the amino acid change is tolerated so scores nearer zero are more likely to be deleterious. The qualitative prediction is derived from this score such that substitutions with a score < 0.05 are called 'deleterious' and all others are called 'tolerated\"\n      Unique Values: \"10\"\n      Name: \"alt_motif\"\n     Description: \"Altered motif\"\n      Unique Values: \"24\"\n      Name: \"disease_xrefs\"\n     Description: \"\"\n      Unique Values: \"1\"\n      Name: \"polyphen_score\"\n     Description: \"The score assigned to a amino acid substitution by a tool called as PolyPhen (=Polymorphism Phenotyping)  which predicts possible impact of an amino acid substitution on the structure and function of a human protein using straightforward physical and comparative considerations. The PolyPhen score ranges from 0.0 (tolerated) to 1.0 (deleterious). Variants with scores of 0.0 are predicted to be benign. Values closer to 1.0 are more confidently predicted to be deleterious\"\n      Unique Values: \"1\"\n      Name: \"polyphen_prediction\"\n     Description: \"\"The PolyPhen prediction values are - greater than 0.908 - \"\"Probably Damaging\"\"; greater than 0.446 and less than or equal to 0.908 - \"\"Possibly Damaging\"\"; less than or equal to 0.446 - \"\"Benign\"\"\"\n      Unique Values: \"1\"\n      Name: \"is_known_glycoprotein\"\n     Description: \"Yes or no flag to denote if a protein is experimentally proven glycoprotein\"\n      Unique Values: \"1\"\n      Name: \"alt_nt\"\n     Description: \"Alternative nucleotide\"\n      Unique Values: \"1\"\n      Name: \"protein_name\"\n     Description: \"Protein name in the UniProtKB\"\n      Unique Values: \"16\"\n      Name: \"disease_evidence_ECO:0000313\"\n     Description: \"The evidence that could be a PMID or a ClinVar ID that associates disease with the variant\"\n      Unique Values: \"1\"\n      Name: \"effect\"\n     Description: \"The effect on glycosylation or N-sequon due to variation at site of glycosylated amino acid.\"\n      Unique Values: \"3\"\n      Name: \"end_aa_pos\"\n     Description: \"End position of the amino acid in the sequence of a protein in UniProtKB\"\n      Unique Values: \"18\"\n      Name: \"mutation_type\"\n     Description: \"Denotes whether mutation was reported with reference to amino acid or nucleotide\"\n      Unique Values: \"1\"\n      Name: \"disease_description\"\n     Description: \"The disease description and/or definition \"\n      Unique Values: \"1\"\n      Name: \"do_id\"\n     Description: \"Disease Ontology identifier\"\n      Unique Values: \"1\"\n      Name: \"evidence_ECO:0000269\"\n     Description: \"A type of experimental evidence that is used in a manual assertion. Usually the values are PMIDs\"\n      Unique Values: \"1\"\n      Name: \"ref_nt\"\n     Description: \"Reference nucleotide\"\n      Unique Values: \"4\"\n      Name: \"disease\"\n     Description: \"The disease name for the cell line\"\n      Unique Values: \"1\"\n      Name: \"alt_aa\"\n     Description: \"Altered amino acid residue\"\n      Unique Values: \"10\"\n      Name: \"cosmic_id\"\n     Description: \"Legacy mutation identifier (COSM) in Catalogue of Somatic Mutations in Cancer (COSMIC) represents existing COSM mutation identifiers. This identifier remains the same between different assemblies (GRCh37 and GRCh38). All the COSM ids at the same genomic location have been collapsed into one representative COSM id. These ids are maintained to help track existing mutations\"\n      Unique Values: \"1\"\n      Name: \"ref_motif\"\n     Description: \"Reference motif\"\n      Unique Values: \"20\"\n      Name: \"xref_key\"\n     Description: \"Source key associated to the  xref_id\"\n      Unique Values: \"2\"\n      Name: \"minor_allelic_frequency\"\n     Description: \"\"\n      Unique Values: \"1\"\n      Name: \"gene_name\"\n     Description: \"\"Official gene name/symbol assigned to the gene. Human gene name/symbol are approved by the HUGO Gene Nomenclature Committee (HGNC) and mouse gene name are approved by the Mouse Genome Informatics (MGI)\"\n      Unique Values: \"17\"\n      Name: \"disease_evidence_ECO:0000269\"\n     Description: \"The PMID evidence for the association of the disease with the variant.\"\n      Unique Values: \"1\"",
    "downloadurl": "/ln2data/releases/data/v-1.11.2/reviewed/rat_protein_sequenceinfo.csv"
}