【原】單細(xì)胞測(cè)序 | 第2期. 下游數(shù)據(jù)質(zhì)控知多少

新用戶4064dVjo 2023-10-27 發(fā)布于北京

展開全文

從上期起我們已經(jīng)進(jìn)入單細(xì)胞領(lǐng)域【第1期. 單細(xì)胞測(cè)序：揭開生命奧秘的鑰匙】,，本期我們就來(lái)聊聊單細(xì)胞數(shù)據(jù)質(zhì)控。在單細(xì)胞懸液的制備過程中,，由于部分細(xì)胞對(duì)分離條件的不適應(yīng),，或是消化條件不夠溫和導(dǎo)致細(xì)胞應(yīng)激或形成碎片，均有可能導(dǎo)致細(xì)胞數(shù)據(jù)無(wú)法反應(yīng)真實(shí)情況,，比如線粒體比例過高,，檢測(cè)基因數(shù)過少/過多，表達(dá)量過低/過高等等,。其次scRNA-seq數(shù)據(jù)相比RNA-seq數(shù)據(jù),，具有更多的噪聲,，即scRNA-seq會(huì)出現(xiàn)許多細(xì)胞中僅有極少的基因表達(dá)，或者某些基因僅在極少的細(xì)胞中表達(dá),，稱dropout events,。如果不進(jìn)行質(zhì)控，不去除低質(zhì)量的細(xì)胞或表達(dá)過低的基因,，都會(huì)使下游分析產(chǎn)生偏差,，使結(jié)果背離真實(shí)的生物學(xué)現(xiàn)象。

一,、方法學(xué)描述

1.細(xì)胞質(zhì)控

①Cells were discarded according to the following criteria: (1) cells that had fewer than 500 genes (UMI > 0); (2) cells that had fewer than 800 UMI or over 8000 UMI; and (3) cells that had more than 10% mitochondrial UMI counts.【PMID: 34601654】

②In brief, for each combination of sample and cluster, an upper and lower bound were set at the 75th percentile plus 1.5 times the interquartile range (IQR) and the 25th percentile minus 1.5 times IQR, respectively, for total UMI, number of genes (n_gene), entropy, and log(n_gene) × entropy.【PMID: 35732739】

③We discarded cells with the number of genes (nFeatureRNA) less than 1000, the molecule to gene ratio (nCount/nFeature) less than 1.2 and the percentage of mitochondrial genes (percent.mt) larger than 5%.【PMID: 37442136】

2.基因質(zhì)控

genes expressed in at least 3 cells were used for further analysis. 【PMID: 32981416】

二,、參數(shù)設(shè)置標(biāo)準(zhǔn)

從上面的方法學(xué)描述可見，質(zhì)控的標(biāo)準(zhǔn)并不是唯一的,，而是由數(shù)據(jù)驅(qū)動(dòng)的,，一般選擇質(zhì)控參數(shù)可由小提琴圖提示，舉一個(gè)Seurat教程中的例子,，如下：