1写在前面

对于只有只有部分重叠的datasets，合并方法我们依然可以采用Seurat、Harmony，rliger包，本期介绍一下Harmony包的用法。🤩

2用到的包

rm(list = ls())
library(Seurat)
library(SeuratDisk)
library(SeuratWrappers)
library(patchwork)
library(harmony)
library(rliger)
library(RColorBrewer)
library(tidyverse)
library(reshape2)
library(ggsci)
library(ggstatsplot)

3示例数据

这里我们提供1个3’ PBMC dataset和1个whole blood dataset。🥰

umi_gz <- gzfile("./GSE149938_umi_matrix.csv.gz",'rt')  
umi <- read.csv(umi_gz,check.names = F,quote = "")

matrix_3p    <- Read10X_h5("./3p_pbmc10k_filt.h5",use.names = T)

创建Seurat对象。🐶

srat_wb <- CreateSeuratObject(t(umi),project = "whole_blood")
srat_3p <- CreateSeuratObject(matrix_3p,project = "pbmc10k_3p")
rm(umi_gz)
rm(umi)
rm(matrix_3p)
srat_wb
srat_3p

4修改metadata

为了方便后续分析，这里我们对metadata进行一下注释修改。

colnames(srat_wb@meta.data)[1] <- "cell_type"
srat_wb@meta.data$orig.ident <- "whole_blood"
srat_wb@meta.data$orig.ident <- as.factor(srat_wb@meta.data$orig.ident)
head(srat_wb[[]])

5初步合并

5.1 简单合并

这里我们先用merge将2个数据集简单合并在一起。（这里我们默认做过初步过滤了哈，具体的大家可以看一下上期的教学。）😘

wb_harmony  <- merge(srat_3p,srat_wb)

5.2 标准操作

我们在这里做一下Normalization，寻找高变基因等等标准操作。👀

wb_harmony <- NormalizeData(wb_harmony, verbose = F)
wb_harmony <- FindVariableFeatures(wb_harmony, selection.method = "vst", nfeatures = 2000, verbose = F)
wb_harmony <- ScaleData(wb_harmony, verbose = F)
wb_harmony <- RunPCA(wb_harmony, npcs = 30, verbose = F)
wb_harmony <- RunUMAP(wb_harmony, reduction = "pca", dims = 1:30, verbose = F)

6harmony合并数据

6.1 合并前

p1 <- DimPlot(object = wb_harmony, reduction = "pca", 
              pt.size = .1, group.by = "orig.ident") + 
  scale_color_npg()+
  NoLegend()

p2 <- VlnPlot(object = wb_harmony, features = "PC_1", 
              group.by = "orig.ident", pt.size = .1) + 
  scale_color_npg()+
  NoLegend()

p1+p2

DimPlot(wb_harmony,reduction = "umap",
        group.by = "orig.ident") + 
  scale_color_npg()+
  plot_annotation(title = "10k 3' PBMC and whole blood, before integration")

6.2 开始合并

wb_harmony <- wb_harmony %>% 
  RunHarmony("orig.ident", plot_convergence = T)

6.3 查看信息

harmony_embeddings <- Embeddings(wb_harmony, 'harmony')
harmony_embeddings[1:5, 1:5]

6.4 可视化-harmony

harmony合并后。

p1 <- DimPlot(object = wb_harmony, reduction = "harmony", pt.size = .1, group.by = "orig.ident") + 
  scale_color_npg()+
  NoLegend()

p2 <- VlnPlot(object = wb_harmony, features = "harmony_1", group.by = "orig.ident", pt.size = .1) + 
  scale_fill_npg()+
  NoLegend()
p1 +p2 

6.5 可视化-UMAP

harmony合并后。

wb_harmony <- SetIdent(wb_harmony,value = "orig.ident")
DimPlot(wb_harmony,reduction = "umap") + 
  scale_color_npg()+
  plot_annotation(title = "10k 3' PBMC and whole blood, after integration (Harmony)")

7降维与聚类

7.1 寻找clusters

wb_harmony <- wb_harmony %>% 
  RunUMAP(reduction = "harmony", dims = 1:30, verbose = F) %>% 
  FindNeighbors(reduction = "harmony", k.param = 10, dims = 1:30) %>% 
  FindClusters() %>% 
  identity()

wb_harmony <- SetIdent(wb_harmony,value = "seurat_clusters")

ncluster <- length(unique(wb_harmony[[]]$seurat_clusters))

mycol <- colorRampPalette(brewer.pal(8, "Set2"))(ncluster)

DimPlot(wb_harmony,label = T,
        cols = mycol, repel = T) + 
  NoLegend()

7.3 具体查看及可视化

我们看下各个clusters在两个datasets各有多少细胞。

count_table <- table(wb_harmony@meta.data$seurat_clusters, wb_harmony@meta.data$orig.ident)
count_table

#### 可视化
count_table %>% 
  as.data.frame() %>% 
  ggbarstats(x = Var2, 
             y = Var1,
             counts = Freq)+
  scale_fill_npg()

需要示例数据的小伙伴，在公众号回复Merge2获取吧！

点个在看吧各位~ ✐.ɴɪᴄᴇ ᴅᴀʏ 〰