背景
受人之託查詢某幾個基因在各種腦細胞中的表達量，我第一時間想到的是上 Allen Brain Map，最後鎖定了 Cell Types Database: RNA-Seq Data 這個數據庫。但是
- 互動熱圖中的各種細胞一個個查看其數據不太方便，光 astrocyte 就分了3種細胞亞型，也許目標細胞還藏在一個個細胞類型 ID 縮寫當中。
- 下載區的表達矩陣文件，csv 有幾十G，不是一般的服務器能夠讀取的；HDF5 文件雖然只有 5 G，使用 R 包
rhdf5的h5read函數讀取，所佔的運行內存和讀取速度也仍然十分可觀，並不是非常經濟的辦法。
於是想到了利用網頁的 API 下載整理爲 Excel 數據以減輕負擔的方法。
API 分析
所有查詢，包括可用的數據集、所有細胞類型和各種基因的表達量數據都通過 POST 方法發送到https://idf-api-prod.aibs-idk-prod.net/獲得查詢結果，請求體是 json 數據，其中
dataset：所查詢的數據集query：GraphQL 檢索式groupBy：結果的排序方式operator：返回的數值的單位
默認的請求體 json 如下：
{
"variables": {
"rows": [
"1_CR",
"2_Meis2",
"3_Meis2",
"4_Meis2 HPF",
"5_Lamp5 Lhx6",
"6_Lamp5 Lhx6",
"7_Lamp5 Lhx6",
"8_Lamp5 Lhx6",
"9_Lamp5 Lhx6",
"10_Lamp5",
"11_Lamp5",
"12_Lamp5",
"13_Lamp5",
"14_Lamp5",
"15_Lamp5",
"16_Lamp5",
"17_Lamp5",
"18_Lamp5",
"19_Pax6",
"20_Pax6",
"21_Sncg",
"22_Sncg",
"23_Sncg",
"24_Sncg",
"25_Sncg",
"26_Ntng1 HPF",
"27_Ntng1 HPF",
"28_Ntng1 HPF",
"29_Ntng1 HPF",
"30_Ntng1 HPF",
"31_Sncg",
"32_Sncg",
"33_Sncg",
"34_Sncg",
"35_Sncg",
"36_Sncg",
"37_Sncg",
"38_Sncg",
"39_Sncg",
"40_Vip",
"41_Vip",
"42_Vip",
"43_Vip",
"44_Vip",
"45_Vip",
"46_Vip",
"47_Vip",
"48_Vip",
"49_Vip",
"50_Vip",
"51_Vip",
"52_Vip",
"53_Vip",
"54_Vip HPF",
"55_Vip HPF",
"56_Vip HPF",
"57_Vip Igfbp6",
"58_Vip Igfbp6",
"59_Vip Igfbp6",
"60_Vip Igfbp6",
"61_Vip Igfbp6",
"62_Vip Igfbp6",
"63_Sst Chodl",
"64_Sst Chodl",
"65_Sst Chodl",
"66_Sst",
"67_Sst",
"68_Sst",
"69_Sst",
"70_Sst",
"71_Sst",
"72_Sst",
"73_Sst",
"74_Sst",
"75_Sst",
"76_Sst",
"77_Sst HPF",
"78_Sst HPF",
"79_Sst",
"80_Sst",
"81_Sst",
"82_Sst",
"83_Sst",
"84_Sst",
"85_Sst",
"86_Sst",
"87_Sst",
"88_Sst",
"89_Sst",
"90_Sst",
"91_Sst",
"92_Sst",
"93_Sst",
"94_Sst",
"95_Sst",
"96_Sst",
"97_Sst",
"98_Sst",
"99_Sst",
"100_Sst",
"101_Sst",
"102_Sst HPF",
"103_Sst HPF",
"104_Sst HPF",
"105_Sst HPF",
"106_Sst HPF",
"107_Sst HPF",
"108_Pvalb",
"109_Pvalb",
"110_Pvalb",
"111_Pvalb",
"112_Pvalb",
"113_Pvalb",
"114_Pvalb",
"115_Pvalb",
"116_Pvalb",
"117_Pvalb",
"118_Pvalb",
"119_Pvalb",
"120_Pvalb",
"121_Pvalb",
"122_Pvalb Vipr2",
"123_Pvalb Vipr2",
"124_L2 IT APr",
"125_L2 IT APr",
"126_L2 IT APr",
"127_L2 IT APr",
"128_L2 IT APr",
"129_L2/3 IT POST-PRE",
"130_L2/3 IT POST-PRE",
"131_L2/3 IT POST-PRE",
"132_L2 IT RSPv-POST-PRE",
"133_L2 IT RSPv-POST-PRE",
"134_L2 IT RSP-ACA",
"135_L3 IT ENTm",
"136_L3 IT ENTm",
"137_L3 IT ENTm",
"138_L3 IT ENTm",
"139_L3 IT ENTl",
"140_L3 IT ENTl",
"141_L2/3 IT PAR",
"142_L2/3 IT PAR",
"143_L2/3 IT PAR",
"144_L2/3 IT PAR",
"145_L2/3 IT PAR",
"146_L2 IT ENTm",
"147_L2 IT ENTm",
"148_L2 IT ENTm",
"149_L2 IT ENTm",
"150_L2 IT ENTm",
"152_L2/3 IT ENTl",
"153_L2/3 IT ENTl",
"154_L2/3 IT ENTl",
"159_L2/3 IT AI",
"158_L2/3 IT AI",
"151_L2 IT ENTl",
"155_L2/3 IT ENTl",
"156_L2/3 IT ENTl",
"157_L2/3 IT ENTl",
"160_L2/3 IT ENTl",
"161_L2/3 IT ENTl",
"162_L2/3 IT CTX",
"163_L2/3 IT CTX",
"164_L2/3 IT CTX",
"165_L2/3 IT CTX",
"166_L2/3 IT CTX",
"167_L2/3 IT CTX",
"168_L2/3 IT CTX",
"169_L2/3 IT CTX",
"170_L2/3 IT CTX",
"171_L2/3 IT CTX",
"172_L2/3 IT ProS",
"173_L2/3 IT ProS",
"174_IT HATA",
"175_IT HATA",
"176_IT HATA",
"177_IT HATA",
"178_L4 IT CTX",
"179_L4 IT CTX",
"180_L4 IT CTX",
"181_L4 IT CTX",
"182_L4/5 IT CTX",
"183_L4/5 IT CTX",
"184_L4/5 IT CTX",
"185_L4/5 IT CTX",
"186_L4/5 IT CTX",
"187_L4/5 IT CTX",
"188_L4/5 IT CTX",
"189_L4/5 IT CTX",
"190_L4/5 IT CTX",
"191_L4/5 IT CTX",
"192_L4/5 IT CTX",
"193_L4/5 IT CTX",
"194_L5 IT RSP-ACA",
"195_L5 IT RSP-ACA",
"196_L5 IT CTX",
"197_L5 IT CTX",
"198_L5 IT CTX",
"199_L5 IT CTX",
"200_L5 IT CTX",
"201_L5 IT CTX",
"202_L5 IT CTX",
"203_L5/6 IT CTX",
"204_L5/6 IT CTX",
"205_L5/6 IT CTX",
"206_L5/6 IT CTX",
"207_L5/6 IT CTX",
"208_L5/6 IT CTX",
"209_L5/6 IT CTX",
"210_L5/6 IT TPE-ENT",
"211_L5/6 IT TPE-ENT",
"212_L5/6 IT TPE-ENT",
"213_L5/6 IT TPE-ENT",
"214_L5/6 IT PFC",
"215_L5/6 IT TPE-ENT",
"216_L5/6 IT TPE-ENT",
"217_L6 IT CTX",
"218_L6 IT CTX",
"219_L6 IT CTX",
"220_L6 IT CTX",
"221_L6 IT CTX",
"222_L6 IT CTX",
"223_L6 IT CTX",
"224_L6 IT CTX",
"225_L6 IT CTX",
"226_L6 IT CTX",
"227_L6 IT CTX",
"228_L6 IT CTX",
"229_L6 IT CTX",
"230_L6 IT CTX",
"231_L6 IT CTX",
"232_L6 IT CTX",
"233_L6 IT ENTl",
"234_L6 IT ENTl",
"235_L6 IT ENTl",
"236_Car3",
"237_Car3",
"238_Car3",
"239_L5 PT CTX",
"240_L5 PT CTX",
"241_L5 PT CTX",
"242_L5 PT CTX",
"243_L5 PT CTX",
"244_L5 PT CTX",
"245_L5 PT CTX",
"246_L5 PT CTX",
"247_L5 PT CTX",
"248_L5 PT CTX",
"249_L5 PT CTX",
"250_L5 PT CTX",
"251_L5 PT CTX",
"252_L5 PT CTX",
"253_L5 PT CTX",
"254_L5 PT CTX",
"255_L5 PT CTX",
"256_L5 PT CTX",
"257_L5 PT CTX",
"258_L5 PT CTX",
"259_L5 PT CTX",
"260_L5 PT CTX",
"261_L4 RSP-ACA",
"262_L4 RSP-ACA",
"263_L5 PPP",
"264_L5/6 NP CTX",
"265_L5/6 NP CTX",
"266_L5/6 NP CTX",
"267_L5/6 NP CTX",
"268_L5/6 NP CTX",
"269_L5/6 NP CTX",
"270_L5/6 NP CT CTX",
"271_NP SUB",
"272_NP SUB",
"273_NP SUB",
"274_NP SUB",
"277_NP PPP",
"275_NP PPP",
"276_NP PPP",
"278_L6 CT CTX",
"279_L6 CT CTX",
"280_L6 CT CTX",
"281_L6 CT CTX",
"282_L6 CT CTX",
"283_L6 CT CTX",
"284_L6 CT CTX",
"285_L6 CT CTX",
"286_L6 CT CTX",
"287_L6 CT CTX",
"288_L6 CT CTX",
"289_L6 CT CTX",
"290_L6 CT CTX",
"291_L6 CT CTX",
"292_L6 CT CTX",
"293_L6 CT CTX",
"294_CT SUB",
"295_CT SUB",
"296_CT SUB",
"297_CT SUB",
"298_L6 CT ENT",
"299_L6 CT ENT",
"308_L6b RHP",
"300_L6b ENT",
"301_L6b ENT",
"302_L6b ENT",
"303_L6b CTX",
"304_L6b CTX",
"305_L6b CTX",
"306_L6b CTX",
"307_L6b CTX",
"310_L6b RHP",
"309_L6b RHP",
"311_L6b CTX",
"312_L6b CTX",
"313_L6b CTX",
"314_L6b CTX",
"315_L6b CTX",
"316_L6b CTX",
"317_L6b CTX",
"318_SUB",
"319_SUB",
"320_SUB",
"321_SUB",
"322_ProS",
"325_ProS",
"323_ProS",
"324_ProS",
"326_ProS",
"327_ProS",
"328_ProS",
"329_CA1-ProS",
"330_CA1-ProS",
"331_CA1-ProS",
"332_CA1-ProS",
"333_CA1-ProS",
"334_CA1-ve",
"335_CA1-ve",
"336_CA1-ve",
"337_CA1",
"338_CA1",
"339_CA1",
"340_CA1",
"341_CA1",
"342_CA1",
"343_CA1",
"344_CA1",
"345_CA1",
"346_CA1-do",
"347_CA1-do",
"348_CA1-do",
"349_Mossy",
"350_Mossy",
"351_CA3-ve",
"352_CA3-ve",
"353_CA3-ve",
"354_CA3-ve",
"355_CA3-ve",
"356_CA3-do",
"357_CA3-do",
"358_CA3-do",
"359_CA2-IG-FC",
"360_CA2-IG-FC",
"361_DG",
"362_DG",
"363_DG",
"364_DG",
"365_Oligo",
"366_Oligo",
"367_Oligo",
"368_Oligo",
"369_Oligo",
"370_Oligo",
"371_Oligo",
"372_Oligo",
"373_Oligo",
"374_Oligo",
"375_Oligo",
"376_Astro",
"377_Astro",
"378_Astro",
"379_Endo",
"380_SMC-Peri",
"381_SMC-Peri",
"382_SMC-Peri",
"383_VLMC",
"384_VLMC",
"385_VLMC",
"386_Micro-PVM",
"387_Micro-PVM",
"388_Micro-PVM"
],
"features": [
"Aqp4",
"C1qa",
"Chodl",
"Cldn5",
"Ctss",
"Cux2",
"Fam84b",
"Fezf2",
"Foxp2",
"Gad1",
"Gad2",
"Lamp5",
"Mbp",
"Ndnf",
"Nxph4",
"Pvalb",
"Rorb",
"Sla2",
"Slc17a7",
"Slc32a1",
"Sncg",
"Sst",
"Sulf1",
"Vip"
],
"dataset": "aibs_mouse_ctx-hpf_10x",
"operator": "TRIMMED_MEANS",
"groupBy": "CLUSTER_LABEL"
},
"query": "query ($dataset: String, $features: [String], $rows: [String], $operator: MatrixAggregationOperator, $groupBy: MatrixAggregationCellMetadata) {\n aggregateRowsOnFeatureMatrix(dataset: $dataset, features: $features, rows: $rows, operator: $operator, groupBy: $groupBy) {\n groupByResults {\n row\n featureResults {\n feature\n value\n __typename\n }\n __typename\n }\n __typename\n }\n}\n"
}
成果
最終代碼上傳到了 Gitee 倉庫。