Seurat分析流程入门
单细胞 RNA 测序(scRNA-seq)基础知识可查看以下文章:
单细胞 RNA 测序(scRNA-seq)工作流程入门
单细胞 RNA 测序(scRNA-seq)细胞分离与扩增
单细胞 RNA 测序(scRNA-seq)SRA 数据下载及 fastq-dumq 数据拆分
单细胞 RNA 测序(scRNA-seq)Cellranger 流程入门和数据质控
单细胞 RNA 测序(scRNA-seq)构建人类参考基因组注释
单细胞RNA测序(scRNA-seq)cellranger count的细胞定量和aggr整合
1. 数据与R包准备
以下代码在RStudio中实现。
1.1 PMBC数据下载
下载2700个10X单细胞-外周血单核细胞(PBMC)数据集。
# Seurat_1.R
################### 1. 数据获取及读取 ###################
# 切换至工作目录
setwd("F:/scRNA-seq/Seurat4.0")
# 下载PMBC数据
download.file("https://cf.10xgenomics.com/samples/cell/pbmc3k/pbmc3k_filtered_gene_bc_matrices.tar.gz",
destfile = "pbmc3k_filtered_gene_bc_matrices.tar.gz")
# 解压
untar(tarfile = "pbmc3k_filtered_gene_bc_matrices.tar.gz", exdir = "./")
# 查看当前目录文件
dir()
# [1] "filtered_gene_bc_matrices" "pbmc3k_filtered_gene_bc_matrices.tar.gz"
# [3] "Seurat_1.R"
# 查看解压目录文件
dir("./filtered_gene_bc_matrices/hg19")
# [1] "barcodes.tsv" "genes.tsv" "matrix.mtx"
1.2 分析R包安装
# 选择清华镜像安装
options("repos"=c(CRAN="https://mirrors.tuna.tsinghua.edu.cn/CRAN/"))
install.packages("BiocManager")
BiocManager::install("dplyr")
BiocManager::install("Seurat")
BiocManager::install("patchwork")
BiocManager::install("ggplot2")
2. 数据预处理
2.1 构建单细胞Seurat对象
Read10X函数读取,返回独特的分子识别(UMI)计数矩阵。 矩阵中的行值表示每个功能(即基因); 列值表示在每个细胞中检测到的分子数量。
################### 2. 数据预处理 ###################
library(dplyr)
library(Seurat)
library(patchwork)
# 读取当前目录 2700个10X单细胞-外周血单核细胞(PBMC)数据集
pbmc.data <- Read10X(data.dir = "./filtered_gene_bc_matrices/hg19")
# 查看部分数据
pbmc.data[1:5, 1:3]
# 5 x 3 sparse Matrix of class "dgCMatrix"
# AAACATACAACCAC-1 AAACATTGAGCTAC-1 AAACATTGATCAGC-1
# MIR1302-10 . . .
# FAM138A . . .
# OR4F5 . . .
# RP11-34P13.7 . . .
# RP11-34P13.8 . . .
# 创建Seurat对象
# 每个基因至少在3个细胞中表达
# 每个细胞至少检测到200个基因
pbmc.obj <- CreateSeuratObject(counts = pbmc.data,
project = "pbmc3k",
min.cell = 3,
min.features = 200)
pbmc.obj
# An object of class Seurat
# 13714 features across 2700 samples within 1 assay
# Active assay: RNA (13714 features, 0 variable features)
# 1 layer present: counts
# 细胞名称重命名
pbmc.obj <- RenameCells(pbmc.obj, add.cell.id = "PBMC")
2.2 单细胞数据质控
为分析得到高质量和可靠的单细胞数据,保证后续分析结果的准确度,通常会设置以下QC指标过滤细胞:
常见QC指标:
- 在每个细胞中检测到的基因数量: 低质量细胞或空液滴通常只有很少的基因; 细胞双倍或多胞可能表现出异常高的基因计数。
- 细胞内检测到的分子总数:(与独特的基因密切相关)
- 细胞中的线粒体基因组的百分比: 低质量/死细胞经常表现出线粒体污染;使用PercentageFeatureSet()函数计算线粒体 QC 指标。
2.2.1 通过细胞counts数和feature数的分布确定过滤条件阈值
head([email]pbmc.obj@meta.data[/email])
# orig.ident nCount_RNA nFeature_RNA percent.mt
# PBMC_AAACATACAACCAC-1 pbmc3k 2419 779 3.0177759
# PBMC_AAACATTGAGCTAC-1 pbmc3k 4903 1352 3.7935958
# PBMC_AAACATTGATCAGC-1 pbmc3k 3147 1129 0.8897363
# PBMC_AAACCGTGCTTCCG-1 pbmc3k 2639 960 1.7430845
# PBMC_AAACCGTGTATGCG-1 pbmc3k 980 521 1.2244898
# PBMC_AAACGCACTGGTAC-1 pbmc3k 2163 781 1.6643551
# 细胞counts数和feature数的分布
VlnPlot(pbmc.obj, features = c("nCount_RNA","nFeature_RNA"))
2.2.2 细胞线粒体基因和和核糖体基因含量获取
# 获取基因名以MT-开头的线粒体基因
pbmc.obj[["percent.mt"]] <- PercentageFeatureSet(pbmc.obj,
pattern = "^MT-")
head([email]pbmc.obj@meta.data[/email], 5)
# orig.ident nCount_RNA nFeature_RNA percent.mt
# PBMC_AAACATACAACCAC-1 PMBC 2419 779 3.0177759
# PBMC_AAACATTGAGCTAC-1 PMBC 4903 1352 3.7935958
# PBMC_AAACATTGATCAGC-1 PMBC 3147 1129 0.8897363
# PBMC_AAACCGTGCTTCCG-1 PMBC 2639 960 1.7430845
# PBMC_AAACCGTGTATGCG-1 PMBC 980 521 1.2244898
# 统计核糖体含量, 核糖体基因命名规则为"RPS"或"RPL"
rRNA.genes <- rownames(pbmc.obj)[grep("^RP[SL]",
rownames(pbmc.obj),
ignore.case=TRUE)]
pbmc.obj$percent.rRNA <- PercentageFeatureSet(pbmc.obj, features = rRNA.genes)
pbmc.obj[["percent.rRNA"]] <- PercentageFeatureSet(pbmc.obj, pattern = "^RP[SL]")
head([email]pbmc.obj@meta.data[/email])
# orig.ident nCount_RNA nFeature_RNA percent.mt percent.rRNA
# PBMC_AAACATACAACCAC-1 PMBC 2419 779 3.0177759 43.69574
# PBMC_AAACATTGAGCTAC-1 PMBC 4903 1352 3.7935958 42.40261
# PBMC_AAACATTGATCAGC-1 PMBC 3147 1129 0.8897363 31.68097
# PBMC_AAACCGTGCTTCCG-1 PMBC 2639 960 1.7430845 24.25161
# PBMC_AAACCGTGTATGCG-1 PMBC 980 521 1.2244898 14.89796
# PBMC_AAACGCACTGGTAC-1 PMBC 2163 781 1.6643551 36.19972
VlnPlot(pbmc.obj, features = c("percent.mt","percent.rRNA"), pt.size = 1)
2.2.3 过滤不符合质控指标的细胞
# 标题显示数字为相关性
p1 <- FeatureScatter(pbmc.obj,
feature1 = "nCount_RNA",
feature2 = "nFeature_RNA")
p2 <- FeatureScatter(pbmc.obj,
feature1 = "nCount_RNA",
feature2 = "percent.mt")
p1 + p2
#相关性为0.95和-0.13
# 使用subset函数过滤seurat对象
# 过滤具有UMI计数超过 2500 或小于 200的细胞, >5%线粒体的细胞
pbmc.obj<- subset(pbmc.obj,
subset = nCount_RNA > 300 & nCount_RNA < 10000 &
nFeature_RNA > 200 & nFeature_RNA < 2500 &
percent.mt < 5)
pbmc.obj
# An object of class Seurat
# 13714 features across 2638 samples within 1 assay
# Active assay: RNA (13714 features, 0 variable features)
# 1 layer present: count
# 过滤后可视化QC指标 "nFeature_RNA", "nCount_RNA", "percent.mt"
VlnPlot(pbmc.obj, features = c("nFeature_RNA","nCount_RNA","percent.mt"),
cols = 3, pt.size = 1)
3 单细胞数据的标准化和均一化
标准化和均一化的处理目的是:
- 使细胞间表达具有可比性
- 使表达量分布符合统计学分布
3.1 数据标准化
最常用的方法是对测序深度进行标准化,使每个细胞具有相同的reads数据量。 Seurat默认的标准化方法为 LogNormalize : 以e为底数,log(每个细胞中基因的nCount_RNA / 该细胞内总Count*10000 + 1)
################### 3. 数据标准化和均一化 ###################
pbmc.obj <- NormalizeData(pbmc.obj,
normalization.method = 'LogNormalize',
scale.factor = 10000)
pbmc.obj@assays$RNA$data[0:5, 0:3]
# 5 x 3 sparse Matrix of class "dgCMatrix"
# PBMC_AAACATACAACCAC-1 PBMC_AAACATTGAGCTAC-1 PBMC_AAACATTGATCAGC-1
# AL627309.1 . . .
# AP006222.2 . . .
# RP11-206L10.2 . . .
# RP11-206L10.9 . . .
# LINC00115 . . .
3.2 筛选高变基因HGVS
高变基因: 计算数据集中显示高变异的特征子集(即在某些细胞中表达强烈,在另一些细胞中表达得很低); 一般使用均值与方差之间的关系来挑选高变基因。
为了减轻下游分析工具的计算负担、减少数据中的噪声并方便数据可视化, **在下游分析中关注这些基因有助于突出单细胞数据集中的生物信号**,。
筛选高变基因方法:
- vst(默认):首先利用loess回归对log(variance)和log(mean)拟合一条直线,然后利用观测均值和期望方差对基因表达量进行标准化,最后根据标准化后的表达量计算方差。
- mean.var.plot(mvp): 首先分别计算每个基因的平均表达量和离散情况,然后根据平均表达量将基因们分散到一定数量(默认是20个)的小区间(bin)中,并且计算每个区间的z-score。
- dispersion (disp): 挑选离差值最高的基因。
# 默认情况下,使用每个数据集的 2,000 个基因
pbmc.obj <- FindVariableFeatures(pbmc.obj,
selection.method = "vst",
nfeatures = 2000)
# 获取前10个高变基因
top10.hgvs <- head(VariableFeatures(pbmc.obj), 10)
# [1] "PPBP" "LYZ" "S100A9" "IGLL5" "GNLY" "FTL" "PF4" "FTH1" "GNG11"
# [10] "S100A8"
p3 <- VariableFeaturePlot(pbmc.obj)
p4 <- LabelPoints(plot = p3, points = top10.hgvs, repel = TRUE)
p3 + p4
3.3 数据均一化
均一化对表达矩阵进行scale处理,scale之后的矩阵每个基因表达均值为0。 经过scale之后,所有基因的表达分布基本一致,有助于后续的降维聚类。
# 均一化
all.genes <- rownames(pbmc.obj)
pbmc.obj <- ScaleData(pbmc.obj, features = all.genes)
pbmc.obj@assays$RNA$scale.data[0:5, 0:3]
# PBMC_AAACATACAACCAC-1 PBMC_AAACATTGAGCTAC-1 PBMC_AAACATTGATCAGC-1
# AL627309.1 -0.05812316 -0.05812316 -0.05812316
# AP006222.2 -0.03357571 -0.03357571 -0.03357571
# RP11-206L10.2 -0.04166819 -0.04166819 -0.04166819
# RP11-206L10.9 -0.03364562 -0.03364562 -0.03364562
# LINC00115 -0.08223981 -0.08223981 -0.08223981
####### 根据需求选择 ######
# 移除细胞周期等因素影响
s.genes <- cc.genes$s.genes
g2m.genes <- cc.genes$g2m.genes
pbmc.obj2 <- CellCycleScoring(pbmc.obj, s.features = s.genes, g2m.features = g2m.genes)
head([email]pbmc.obj2@meta.data[/email])
# orig.ident nCount_RNA nFeature_RNA percent.mt percent.rRNA
# PBMC_AAACATACAACCAC-1 PMBC 2419 779 3.0177759 43.69574
# PBMC_AAACATTGAGCTAC-1 PMBC 4903 1352 3.7935958 42.40261
# PBMC_AAACATTGATCAGC-1 PMBC 3147 1129 0.8897363 31.68097
# PBMC_AAACCGTGCTTCCG-1 PMBC 2639 960 1.7430845 24.25161
# PBMC_AAACCGTGTATGCG-1 PMBC 980 521 1.2244898 14.89796
# PBMC_AAACGCACTGGTAC-1 PMBC 2163 781 1.6643551 36.19972
# S.Score G2M.Score Phase
# PBMC_AAACATACAACCAC-1 0.09853841 -0.044716507 S
# PBMC_AAACATTGAGCTAC-1 -0.02364305 -0.046889929 G1
# PBMC_AAACATTGATCAGC-1 -0.02177266 0.074841537 G2M
# PBMC_AAACCGTGCTTCCG-1 0.03794398 0.006575446 S
# PBMC_AAACCGTGTATGCG-1 -0.03309970 0.027910063 G2M
# PBMC_AAACGCACTGGTAC-1 -0.04814181 -0.078164839 G1
4 单细胞数据降维
降维是通过算法最优地保留原始数据的一些关键属性来将数据投影到更低的维度空间中,便于可视化和后续分析。
4.1 PCA降维分析
PCA降维时, Secuat包提供VizDimLoadings(), DimPlot() 和 DimHeatmap() 函数用于单细胞 cells and features 的可视化。
################### 4. 数据降维 ###################
hgvs <- VariableFeatures(object = pbmc.obj)
# PCA分析
pbmc.obj <- RunPCA(pbmc.obj, features = hgvs)
# 打印部分PCA结果
print(pbmc.obj[["pca"]], dims = 1:3, nfeatures = 5)
# PC_ 1
# Positive: CST3, TYROBP, LST1, AIF1, FTL
# Negative: MALAT1, LTB, IL32, IL7R, CD2
# PC_ 2
# Positive: CD79A, MS4A1, TCL1A, HLA-DQA1, HLA-DQB1
# Negative: NKG7, PRF1, CST7, GZMB, GZMA
# PC_ 3
# Positive: HLA-DQA1, CD79A, CD79B, HLA-DQB1, HLA-DPB1
# Negative: PPBP, PF4, SDPR, SPARC, GNG11
# 不同函数的可视化结果
VizDimLoadings(pbmc.obj, dims = 1:2 , reduction = "pca")
DimPlot(pbmc.obj, reduction = "pca", pt.size = 1)
DimHeatmap(pbmc.obj, dims = 1:6 , reduction = "pca", balanced = TRUE)
4.2 选取合适的维度
维度的选择标准是在保证足够的信息同时减少噪音。
# 展示维度信息, 选取20较为合适
ElbowPlot(pbmc.obj)
# 确定主成分个数,耗时长
pbmc.obj <- JackStraw(pbmc.obj, num.replicate = 100)
pbmc.obj <- ScoreJackStraw(pbmc.obj, dims = 1:20)
# 可视化
JackStrawPlot(pbmc.obj, dims = 1:20)
5. 单细胞聚类
基于细胞基因表达谱的相似性,将细胞聚类成簇。通常是任何单细胞分析的第一个中间结果,细胞聚类可以帮助我们推断数据集中各细胞的类别。
5.1 聚类分簇
表达谱的相似性采用降维之后表达空间上的欧氏距离度量。
################### 5. 数据聚类 ###################
# 单细胞聚类
# 最近邻算法对细胞进行聚类
pbmc.obj <- FindNeighbors(pbmc.obj, reduction = "pca", dims = 1:10)
# 使用多分辨率模块优化算法,迭代将细胞进行分类
pbmc.obj <- FindClusters(pbmc.obj, resolution = 0.5)
head([email]pbmc.obj@meta.data[/email])
# orig.ident nCount_RNA nFeature_RNA percent.mt percent.rRNA
# PBMC_AAACATACAACCAC-1 PMBC 2419 779 3.0177759 43.69574
# PBMC_AAACATTGAGCTAC-1 PMBC 4903 1352 3.7935958 42.40261
# PBMC_AAACATTGATCAGC-1 PMBC 3147 1129 0.8897363 31.68097
# PBMC_AAACCGTGCTTCCG-1 PMBC 2639 960 1.7430845 24.25161
# PBMC_AAACCGTGTATGCG-1 PMB 980 521 1.2244898 14.89796
# PBMC_AAACGCACTGGTAC-1 PMBC 2163 781 1.6643551 36.19972
# RNA_snn_res.0.5 seurat_clusters
# PBMC_AAACATACAACCAC-1 2 2
# PBMC_AAACATTGAGCTAC-1 3 3
# PBMC_AAACATTGATCAGC-1 2 2
# PBMC_AAACCGTGCTTCCG-1 1 1
# PBMC_AAACCGTGTATGCG-1 6 6
# PBMC_AAACGCACTGGTAC-1 2 2
table([email]pbmc.obj@meta.data[/email]$seurat_clusters)
# 0 1 2 3 4 5 6 7 8
# 711 480 472 344 279 162 144 32 14
5.2 UMAP 可视化
Seurat R包提供UMAP和t-SNE两种降维可视化算法。
# UMAP聚类可视化
pbmc.obj <- RunUMAP(pbmc.obj, reduction = "pca",
dims = 1:10, verbose = TRUE)
DimPlot(pbmc.obj, reduction = "umap",
label = TRUE, pt.size = 1.5)
head(pbmc.obj@reductions$[email]umap@cell.embeddings[/email])
# umap_1 umap_2
# PBMC_AAACATACAACCAC-1 -4.611246 -3.0574470
# PBMC_AAACATTGAGCTAC-1 -2.061398 9.7727327
# PBMC_AAACATTGATCAGC-1 -5.386918 -6.3924266
# PBMC_AAACCGTGCTTCCG-1 10.931999 3.5166690
# PBMC_AAACCGTGTATGCG-1 -9.780782 0.6403372
# PBMC_AAACGCACTGGTAC-1 -3.459144 -4.8853937
5.3 tSNE 可视化
# tSNE聚类可视化
pbmc.obj <- RunTSNE(pbmc.obj, reduction = "pca", dims = 1:10)
DimPlot(pbmc.obj, reduction = "tsne",
label = TRUE, pt.size = 1.5)
# 自定义颜色, 聚类靠前颜色显著
# DimPlot(pbmc.obj, reduction = "tsne", cols = c("red", "blue", 3:10),
# label = TRUE, pt.size = 1.5)
head(pbmc.obj@reductions$[email]tsne@cell.embeddings[/email])
# tSNE_1 tSNE_2
# PBMC_AAACATACAACCAC-1 -11.701490 7.120466
# PBMC_AAACATTGAGCTAC-1 -21.981401 -21.330793
# PBMC_AAACATTGATCAGC-1 -1.292978 23.822324
# PBMC_AAACCGTGCTTCCG-1 30.877776 -9.926240
# PBMC_AAACCGTGTATGCG-1 -34.619197 8.773753
# PBMC_AAACGCACTGGTAC-1 -3.046157 13.013471
自定义颜色可视化图
红色和蓝色为自定义颜色。
6. 单细胞聚类簇鉴定
在基因水平上,对每个簇的marker gene进行分析,这些marker gene包含细胞簇的特征,可帮助根据细胞簇生物学功能等特性进行分类。
6.1 单细胞差异基因分析
min.pct: 在2组细胞簇中检测特征的最小百分比。
# 单细胞聚簇分类
# 细胞簇差异基因(DEGs)分析
cluster1.marker <- FindMarkers(pbmc.obj, ident.1 = 0,
min.pct = 0.1, logfc.threshold = 0.25)
head(cluster1.marker)
# p_val avg_log2FC pct.1 pct.2 p_val_adj
# RPS12 1.806317e-144 0.7439459 1.000 0.991 2.477183e-140
# RPS6 7.135900e-142 0.6862148 1.000 0.995 9.786173e-138
# RPS27 5.257820e-140 0.7298479 0.999 0.992 7.210575e-136
# RPL32 4.229582e-136 0.6184804 0.999 0.995 5.800448e-132
# RPS14 1.799019e-130 0.6283021 1.000 0.994 2.467175e-126
# CYBA 1.235790e-129 -1.7444394 0.658 0.917 1.694762e-125
# 从cluster0到cluster3中区分cluster5 marker gene
cluster5.marker <- FindMarkers(pbmc.obj, ident.1 = 5,
ident.2 = c(0,3), min.pct = 0.25)
head(cluster5.marker)
# p_val avg_log2FC pct.1 pct.2 p_val_adj
# FCGR3A 2.150929e-209 6.832372 0.975 0.039 2.949784e-205
# IFITM3 6.103366e-199 6.181000 0.975 0.048 8.370156e-195
# CFD 8.891428e-198 6.052575 0.938 0.037 1.219370e-193
# CD68 2.374425e-194 5.493138 0.926 0.035 3.256286e-190
# RP11-290F20.3 9.308287e-191 6.335402 0.840 0.016 1.276538e-186
# IFI30 1.575100e-181 6.382751 0.796 0.014 2.160093e-17
# 比较剩余的细胞簇找到marker gene
# only.pos为TRUE表示仅记录阳性结果
all.marker <- FindAllMarkers(pbmc.obj, only.pos = TRUE,
min.pct = 0.1 , logfc.threshold = 0.25)
head(all.marker)
# p_val avg_log2FC pct.1 pct.2 p_val_adj cluster gene
# RPS12 1.806317e-144 0.7439459 1.000 0.991 2.477183e-140 0 RPS12
# RPS6 7.135900e-142 0.6862148 1.000 0.995 9.786173e-138 0 RPS6
# RPS27 5.257820e-140 0.7298479 0.999 0.992 7.210575e-136 0 RPS27
# RPL32 4.229582e-136 0.6184804 0.999 0.995 5.800448e-132 0 RPL32
# RPS14 1.799019e-130 0.6283021 1.000 0.994 2.467175e-126 0 RPS14
# RPS25 5.507298e-123 0.7635765 0.997 0.975 7.552709e-119 0 RPS25
6.2 mark gene可视化
mark gene为已知的特定细胞中特异性表达的基因,根据已知的数据库(如CellMarker、PanglaoDB等)和发表的文献。
与mark gene可视化相关函数:
- VlnPlot() (shows expression probability distributions across clusters)
- FeaturePlot() (visualizes feature expression on a tSNE or PCA plot) are our most commonly used visualizations.
- DoHeatmap() generates an expression heatmap for given cells and features.
- RidgePlot(), CellScatter(), and DotPlot() as additional methods to view your dataset.
library(dplyr)
library(ggplot2)
# marker基因可视化
marker.show <- c("MS4A1", "GNLY", "CD3E", "CD14", "FCER1A",
"FCGR3A", "LYZ", "PPBP", "CD8A")
VlnPlot(pbmc.obj, features = marker.show)
VlnPlot(pbmc.obj, features = marker.show, slot = "counts", log = TRUE)
FeaturePlot(pbmc.obj, features = marker.show)
# 根据中cluster列分组,选取按avg_log2FC排序
top10.marker <- all.marker %>% group_by(cluster) %>% top_n(n=10, wt = avg_log2FC)
DoHeatmap(pbmc.obj, features = top10.marker$ge) + NoLegend()
# 点图, 颜色深浅表示表达量的高低,点大小表示表达量占比
DotPlot(pbmc.obj, features = marker.show) +
theme(axis.text.x = element_text(angle=45, hjust=1))
# 山脊图
RidgePlot(pbmc.obj, features = marker.show)
6.3 细胞类型注释
根据细胞中特异性表达的marker gene来注释细胞簇的类型。
使用marker来轻松地将聚类与已知的单元类型进行匹配:
cluster.cell <- c("Naive CD4 T", "Memory CD4 T", "CD14+ Mono", "B",
"CD8 T", "FCGR3A+ Mono", "NK", "DC", "Platelet")
names(cluster.cell) <- levels(pbmc.obj)
cluster.cell
# 0 1 2 3 4 5
# "Naive CD4 T" "Memory CD4 T" "CD14+ Mono" "B" "CD8 T" "FCGR3A+ Mono"
# 6 7 8
# "NK" "DC" "Platelet"
# 重命名Idents
pbmc.obj <- RenameIdents(pbmc.obj, cluster.cell)
# 可视化
DimPlot(pbmc.obj, reduction = "umap", label=TRUE, pt.size=1)
# 保存为.RDS结果
saveRDS(pbmc.obj, "pbmc.final.RDS")
head([email]pbmc.obj@active.ident[/email])
# PBMC_AAACATACAACCAC-1 PBMC_AAACATTGAGCTAC-1 PBMC_AAACATTGATCAGC-1 PBMC_AAACCGTGCTTCCG-1
# CD14+ Mono B CD14+ Mono Memory CD4 T
# PBMC_AAACCGTGTATGCG-1 PBMC_AAACGCACTGGTAC-1
# NK CD14+ Mono
# Levels: Naive CD4 T Memory CD4 T CD14+ Mono B CD8 T FCGR3A+ Mono NK DC Platelet