Here, Harmony is executed on the Seurat object for Study 2 (Count) data. The parameters and commands are derived from the Harmony documentation.

source("~/Supplemental_code/Data_integration.R")
library(Seurat)
library(dplyr)
library(harmony)
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)
harmony_obj_study2 <- CreateSeuratObject(counts = Stydy2Data, project = "Harmony_study2", min.cells = 5,meta.data = Stydy2Datametadata) %>%
  Seurat::NormalizeData(verbose = FALSE) %>%
  FindVariableFeatures(selection.method = "vst", nfeatures = 2000) %>% 
  ScaleData(verbose = FALSE) %>% 
  RunPCA(pc.genes = pbmc@var.genes, npcs = 20, verbose = FALSE)

Run RunHarmony Seurat wrapper

harmony_obj_study2 = harmony::RunHarmony(harmony_obj_study2,group.by.vars = "Data_id")
harmony_embeddings <- Embeddings(harmony_obj_study2, 'harmony')
harmony_embeddings[1:5, 1:5]
##                     harmony_1 harmony_2 harmony_3 harmony_4 harmony_5
## 1850055018_CS13_S2   18.29675  11.71153 1.0098857  2.911585 0.4413637
## 1850055018_CS32_S12  18.94523  12.48664 0.9780357  2.186623 0.4185233
## 1850055018_CS24_S7   19.57142  13.25548 1.6422007  3.069661 0.1125002
## 1850055018_CS18_S5   20.26635  14.11057 1.6570702  3.608592 0.4950125
## 1850055018_CS25_S8   19.50240  13.85487 1.8494443  3.048712 0.2196713
harmony_obj_study2 <- harmony_obj_study2 %>% 
  RunUMAP(reduction = "harmony", dims = 1:20) %>% 
  FindNeighbors(reduction = "harmony", dims = 1:20) %>% 
  FindClusters(resolution = 0.5) %>% 
  identity()
## Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck
## 
## Number of nodes: 1480
## Number of edges: 54152
## 
## Running Louvain algorithm...
## Maximum modularity in 10 random starts: 0.8502
## Number of communities: 9
## Elapsed time: 0 seconds

visualization colorkey

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 plot

DimPlot(object = harmony_obj_study2, reduction = "pca", pt.size = 1,group.by = "Data_id")+scale_color_manual(values=ColorKeyDataID)

DimPlot(object = harmony_obj_study2, reduction = "pca", pt.size = 1)+scale_color_manual(values=ColorKeyDataID)

Harmony plot

DimPlot(object = harmony_obj_study2, reduction = "harmony", pt.size = 1,group.by = "Data_id") +scale_color_manual(values=ColorKeyDataID)

DimPlot(object = harmony_obj_study2, reduction = "harmony", pt.size = 1) +scale_color_manual(values=ColorKeyDataID)

UMAP plot

DimPlot(harmony_obj_study2, reduction = "umap", pt.size =1,group.by = "Data_id")+scale_color_manual(values=ColorKeyDataID)

DimPlot(harmony_obj_study2, reduction = "umap", pt.size =1)+scale_color_manual(values=ColorKeyDataID)

barplot

harmony_df_study2 = harmony_obj_study2@meta.data 

ggplot(harmony_df_study2, aes(x=seurat_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(harmony_df_study2$seurat_clusters,harmony_df_study2$Class)
## [1] 0.08404987
aricode::NMI(harmony_df_study2$seurat_clusters,harmony_df_study2$Class)
## [1] 0.1380744
ClusterPurity <- function(clusters, classes) {
  sum(apply(table(classes, clusters), 2, max)) / length(clusters)
}
ClusterPurity(harmony_df_study2$seurat_clusters,harmony_df_study2$Class)
## [1] 0.7844595