Seurat pipeline for Study 2 Count data. We combined all datasets of Study 2 into single matrix on the basis of common genes.

source("~/Supplemental_code/Data_integration.R")
library(Seurat)
library(dplyr)
library(ggplot2)

Load datasets

#Create Expression data list
load("~/Supplemental_code/unCTC_datasets/Poonia_et_al._CountData.RData")
load("~/Supplemental_code/unCTC_datasets/Ding_et_al._WBC1_CountData.RData")
load("~/Supplemental_code/unCTC_datasets/Ding_et_al._WBC2_CountData.RData")
load("~/Supplemental_code/unCTC_datasets/Ebright_et_al._CountData.RData")

Make metaData file

Poonia_et_al._CountmetaData = as.data.frame(rep("Poonia_et_al._CountData",ncol(Poonia_et_al._CountData)))
colnames(Poonia_et_al._CountmetaData) = "Data_id"
rownames(Poonia_et_al._CountmetaData) = colnames(Poonia_et_al._CountData)
Poonia_et_al._CountmetaData$Class = "CTC"

Ebright_et_al._CountmetaData = as.data.frame(rep("Ebright_et_al._CountData",ncol(Ebright_et_al._CountData)))
colnames(Ebright_et_al._CountmetaData) = "Data_id"
rownames(Ebright_et_al._CountmetaData) = colnames(Ebright_et_al._CountData)
Ebright_et_al._CountmetaData$Class = "CTC"

Ding_et_al._WBC1_CountmetaData= as.data.frame(rep("Ding_et_al._WBC1_CountData",ncol(Ding_et_al._WBC1_CountData)))
colnames(Ding_et_al._WBC1_CountmetaData) = "Data_id"
rownames(Ding_et_al._WBC1_CountmetaData) = colnames(Ding_et_al._WBC1_CountData)
Ding_et_al._WBC1_CountmetaData$Class = "WBC"

Ding_et_al._WBC2_CountmetaData =as.data.frame(rep("Ding_et_al._WBC2_CountData",ncol(Ding_et_al._WBC2_CountData)))
colnames(Ding_et_al._WBC2_CountmetaData) = "Data_id"
rownames(Ding_et_al._WBC2_CountmetaData) = colnames(Ding_et_al._WBC2_CountData)
Ding_et_al._WBC2_CountmetaData$Class = "WBC"

Data integration based on common genes

Stydy2Data = Data_integration(data_list=list(Poonia_et_al._CountData,
                                             Ebright_et_al._CountData,
                                             Ding_et_al._WBC1_CountData,
                                             Ding_et_al._WBC2_CountData))

Stydy2Datametadata = rbind(Poonia_et_al._CountmetaData,Ebright_et_al._CountmetaData,
                           Ding_et_al._WBC1_CountmetaData,Ding_et_al._WBC2_CountmetaData)
# Create seurat object
Seurat_obj_study2 <- CreateSeuratObject(counts = Stydy2Data, project = "Study_2_data",meta.data = Stydy2Datametadata)
# normalize and identify variable features for each dataset independently
Seurat_obj_study2 <- NormalizeData(Seurat_obj_study2, normalization.method = "LogNormalize", scale.factor = 10000)
#Identification of highly variable features (feature selection)
Seurat_obj_study2 <- FindVariableFeatures(Seurat_obj_study2, selection.method = "vst", nfeatures = 2000)
#Scaling
all.genes <- rownames(Seurat_obj_study2)
Seurat_obj_study2 <- ScaleData(Seurat_obj_study2, features = all.genes)
#perform linear dimensional reduction
Seurat_obj_study2 <- RunPCA(Seurat_obj_study2, features = VariableFeatures(object = Seurat_obj_study2))
#Perform clustering 
Seurat_obj_study2 <- FindNeighbors(Seurat_obj_study2, dims = 1:10)
Seurat_obj_study2 <- FindClusters(Seurat_obj_study2, resolution = 0.5)
## Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck
## 
## Number of nodes: 1480
## Number of edges: 42904
## 
## Running Louvain algorithm...
## Maximum modularity in 10 random starts: 0.9003
## Number of communities: 10
## Elapsed time: 0 seconds
Seurat_obj_study2 <- RunUMAP(Seurat_obj_study2, dims = 1:10)

Colorkey for Visualization

# Visualization
ColorKeyDataID = c("peru","steelblue","darkolivegreen4","palevioletred4",
                   "darkcyan","darkorchid4","darkslategray","firebrick1",
                   "salmon3","paleturquoise1","mediumaquamarine",
                   "greenyellow","black","deepskyblue3","mediumblue",
                   "darkred","gold","gray50","hotpink","khaki3",
                   "yellow4","lavender","cornsilk4","orchid4",
                   "yellow3", "darkgreen","skyblue1","khaki4",
                   "tan4","pink")

PCA

DimPlot(Seurat_obj_study2, reduction = "pca",pt.size =1, group.by = "Data_id")+scale_color_manual(values=ColorKeyDataID)

DimPlot(Seurat_obj_study2, reduction = "pca",pt.size =1, label = FALSE, repel = TRUE)+scale_color_manual(values=ColorKeyDataID)

UMAP

DimPlot(Seurat_obj_study2, reduction = "umap",pt.size =1, group.by = "Data_id")+scale_color_manual(values=ColorKeyDataID)

DimPlot(Seurat_obj_study2, reduction = "umap",pt.size =1, label = FALSE, repel = TRUE)+scale_color_manual(values=ColorKeyDataID)

barplot

seurat_umap_study2  = data.frame(Seurat_obj_study2@meta.data)
seurat_umap_study2$Clusters = Seurat_obj_study2@meta.data$seurat_clusters

ggplot(seurat_umap_study2, aes(x=Clusters, fill = Class)) + theme_classic()+
       geom_bar(stat="count")+scale_color_manual()+
       scale_fill_manual(values = c("dodgerblue4","firebrick3","darkgreen","dark turquoise"))+
       theme(legend.text = element_text(size=14),
       plot.title = element_text(size=16),
       legend.title=element_text(size=20),axis.text=element_text(size=20),
       axis.title=element_text(size=22,face="bold")) +
       guides(colour = guide_legend(override.aes = list(size = 6)))

ARI, NMI and Cluster purity

aricode::ARI(seurat_umap_study2$seurat_clusters,seurat_umap_study2$Class)
## [1] 0.3138856
aricode::NMI(seurat_umap_study2$seurat_clusters,seurat_umap_study2$Class)
## [1] 0.2772147
ClusterPurity <- function(clusters, classes) {
  sum(apply(table(classes, clusters), 2, max)) / length(clusters)
}
ClusterPurity(seurat_umap_study2$seurat_clusters,seurat_umap_study2$Class)
## [1] 0.9628378