Merged All samples with PBMC_10x

1. load libraries

Loading required package: SeuratObject

Loading required package: sp

Attaching package: 'SeuratObject'

The following objects are masked from 'package:base':

    intersect, t

── Installed datasets ──────────────────────────────── SeuratData v0.2.2.9001 ──

✔ pbmcref 1.0.0                         ✔ pbmcsca 3.0.0

────────────────────────────────────── Key ─────────────────────────────────────

✔ Dataset loaded successfully
❯ Dataset built with a newer version of Seurat than installed
❓ Unknown version of Seurat installed

Attaching package: 'dplyr'

The following objects are masked from 'package:stats':

    filter, lag

The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ forcats   1.0.0     ✔ readr     2.1.5
✔ ggplot2   3.5.1     ✔ stringr   1.5.1
✔ lubridate 1.9.3     ✔ tibble    3.2.1
✔ purrr     1.0.2     ✔ tidyr     1.3.1
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

Attaching package: 'magrittr'


The following object is masked from 'package:purrr':

    set_names


The following object is masked from 'package:tidyr':

    extract



Attaching package: 'dbplyr'


The following objects are masked from 'package:dplyr':

    ident, sql


Registered S3 method overwritten by 'SeuratDisk':
  method            from  
  as.sparse.H5Group Seurat



Attaching shinyBS

Loading required package: ggraph


Attaching package: 'ggraph'


The following object is masked from 'package:sp':

    geometry

2. Load Seurat Object

#Load Seurat Object merged from cell lines and a control(PBMC) after filtration
load("../../../0-IMP-OBJECTS/All_Samples_Merged_with_10x_Azitmuth_Annotated.robj")

All_samples_Merged

An object of class Seurat 
36752 features across 59355 samples within 5 assays 
Active assay: RNA (36601 features, 0 variable features)
 2 layers present: data, counts
 4 other assays present: ADT, prediction.score.celltype.l1, prediction.score.celltype.l2, prediction.score.celltype.l3
 2 dimensional reductions calculated: integrated_dr, ref.umap

Summarizing Seurat Object

# Load necessary libraries
library(Seurat)

# Display basic metadata summary
head(All_samples_Merged@meta.data)

# Check if columns such as `orig.ident`, `nCount_RNA`, `nFeature_RNA`, `nUMI`, `ngene`, and any other necessary columns exist
required_columns <- c("orig.ident", "nCount_RNA", "nFeature_RNA", "nUMI", "ngene")
missing_columns <- setdiff(required_columns, colnames(All_samples_Merged@meta.data))

if (length(missing_columns) > 0) {
    cat("Missing columns:", paste(missing_columns, collapse = ", "), "\n")
} else {
    cat("All required columns are present.\n")
}

All required columns are present.

# Check cell counts and features
cat("Number of cells:", ncol(All_samples_Merged), "\n")

Number of cells: 59355 

cat("Number of features:", nrow(All_samples_Merged), "\n")

Number of features: 36601 

# Verify that each `orig.ident` label has the correct number of cells
cat("Cell counts per group:\n")

Cell counts per group:

print(table(All_samples_Merged$orig.ident))

     L1      L2      L3      L4      L5      L6      L7    PBMC PBMC10x 
   5825    5935    6428    6150    6022    5148    5331    8354   10162 

# Check that the cell IDs are unique (which ensures no issues from merging)
if (any(duplicated(colnames(All_samples_Merged)))) {
    cat("Warning: There are duplicated cell IDs.\n")
} else {
    cat("Cell IDs are unique.\n")
}

Cell IDs are unique.

# Check the assay consistency for RNA
DefaultAssay(All_samples_Merged) <- "RNA"

# Check dimensions of the RNA counts layer using the new method
cat("Dimensions of the RNA counts layer:", dim(GetAssayData(All_samples_Merged, layer = "counts")), "\n")

Dimensions of the RNA counts layer: 36601 59355 

cat("Dimensions of the RNA data layer:", dim(GetAssayData(All_samples_Merged, layer = "data")), "\n")

Dimensions of the RNA data layer: 36601 59355 

# Check the ADT assay (optional)
if ("ADT" %in% names(All_samples_Merged@assays)) {
    cat("ADT assay is present.\n")
    cat("Dimensions of the ADT counts layer:", dim(GetAssayData(All_samples_Merged, assay = "ADT", layer = "counts")), "\n")
} else {
    cat("ADT assay is not present.\n")
}

ADT assay is present.
Dimensions of the ADT counts layer: 56 59355 

Azimuth Annotation

# InstallData("pbmcref")
# 
# # The RunAzimuth function can take a Seurat object as input
# All_samples_Merged <- RunAzimuth(All_samples_Merged, reference = "pbmcref")

3. QC

# Remove the percent.mito column
All_samples_Merged$percent.mito <- NULL


# Set identity classes to an existing column in meta data
Idents(object = All_samples_Merged) <- "cell_line"

All_samples_Merged[["percent.rb"]] <- PercentageFeatureSet(All_samples_Merged, 
                                                           pattern = "^RP[SL]")
# Convert 'percent.mt' to numeric, replacing "NaN" with 0
All_samples_Merged$percent.rb <- replace(as.numeric(All_samples_Merged$percent.rb), is.na(All_samples_Merged$percent.rb), 0)



# The [[ operator can add columns to object metadata. This is a great place to stash QC stats
All_samples_Merged[["percent.mt"]] <- PercentageFeatureSet(All_samples_Merged, pattern = "^MT-")

# Convert 'percent.mt' to numeric, replacing "NaN" with 0
All_samples_Merged$percent.mt <- replace(as.numeric(All_samples_Merged$percent.mt), is.na(All_samples_Merged$percent.mt), 0)





VlnPlot(All_samples_Merged, features = c("nFeature_RNA", 
                                         "nCount_RNA", 
                                         "percent.mt",
                                         "percent.rb"), 
                            ncol = 4, pt.size = 0.1) & 
              theme(plot.title = element_text(size=10))

FeatureScatter(All_samples_Merged, feature1 = "percent.mt", 
                                  feature2 = "percent.rb")

VlnPlot(All_samples_Merged, features = c("nFeature_RNA", 
                                    "nCount_RNA", 
                                    "percent.mt"), 
                                      ncol = 3)

FeatureScatter(All_samples_Merged, 
               feature1 = "percent.mt", 
               feature2 = "percent.rb") +
        geom_smooth(method = 'lm')

`geom_smooth()` using formula = 'y ~ x'

FeatureScatter(All_samples_Merged, 
               feature1 = "nCount_RNA", 
               feature2 = "nFeature_RNA") +
        geom_smooth(method = 'lm')

`geom_smooth()` using formula = 'y ~ x'

##FeatureScatter is typically used to visualize feature-feature relationships ##for anything calculated by the object, ##i.e. columns in object metadata, PC scores etc.

FeatureScatter(All_samples_Merged, 
               feature1 = "nCount_RNA", 
               feature2 = "percent.mt")+
  geom_smooth(method = 'lm')

`geom_smooth()` using formula = 'y ~ x'

FeatureScatter(All_samples_Merged, 
               feature1 = "nCount_RNA", 
               feature2 = "nFeature_RNA")+
  geom_smooth(method = 'lm')

`geom_smooth()` using formula = 'y ~ x'

Assign Cell-Cycle Scores

Running SCTransform on assay: RNA

Running SCTransform on layer: counts

vst.flavor='v2' set. Using model with fixed slope and excluding poisson genes.

Variance stabilizing transformation of count matrix of size 27417 by 59355

Model formula is y ~ log_umi

Get Negative Binomial regression parameters per gene

Using 2000 genes, 5000 cells

Found 453 outliers - those will be ignored in fitting/regularization step

Second step: Get residuals using fitted parameters for 27417 genes

Computing corrected count matrix for 27417 genes

Calculating gene attributes

Wall clock passed: Time difference of 10.66191 mins

Determine variable features

Getting residuals for block 1(of 12) for counts dataset

Getting residuals for block 2(of 12) for counts dataset

Getting residuals for block 3(of 12) for counts dataset

Getting residuals for block 4(of 12) for counts dataset

Getting residuals for block 5(of 12) for counts dataset

Getting residuals for block 6(of 12) for counts dataset

Getting residuals for block 7(of 12) for counts dataset

Getting residuals for block 8(of 12) for counts dataset

Getting residuals for block 9(of 12) for counts dataset

Getting residuals for block 10(of 12) for counts dataset

Getting residuals for block 11(of 12) for counts dataset

Getting residuals for block 12(of 12) for counts dataset

Finished calculating residuals for counts

Set default assay to SCT

Warning: The following features are not present in the object: MLF1IP, not
searching for symbol synonyms

Warning: The following features are not present in the object: FAM64A, HN1, not
searching for symbol synonyms

4. Normalize data

# Apply SCTransform
All_samples_Merged <- SCTransform(All_samples_Merged, 
                                  vars.to.regress = c("percent.rb","percent.mt", "CC.Difference", "cell_line"), 
                                  do.scale=TRUE, 
                                  do.center=TRUE, 
                                  verbose = TRUE)

Running SCTransform on assay: RNA

Running SCTransform on layer: counts

vst.flavor='v2' set. Using model with fixed slope and excluding poisson genes.

Variance stabilizing transformation of count matrix of size 27417 by 59355

Model formula is y ~ log_umi

Get Negative Binomial regression parameters per gene

Using 2000 genes, 5000 cells

Found 453 outliers - those will be ignored in fitting/regularization step

Second step: Get residuals using fitted parameters for 27417 genes

Computing corrected count matrix for 27417 genes

Calculating gene attributes

Wall clock passed: Time difference of 8.084556 mins

Determine variable features

Regressing out percent.rb, percent.mt, CC.Difference, cell_line

Centering and scaling data matrix

Getting residuals for block 1(of 12) for counts dataset

Getting residuals for block 2(of 12) for counts dataset

Getting residuals for block 3(of 12) for counts dataset

Getting residuals for block 4(of 12) for counts dataset

Getting residuals for block 5(of 12) for counts dataset

Getting residuals for block 6(of 12) for counts dataset

Getting residuals for block 7(of 12) for counts dataset

Getting residuals for block 8(of 12) for counts dataset

Getting residuals for block 9(of 12) for counts dataset

Getting residuals for block 10(of 12) for counts dataset

Getting residuals for block 11(of 12) for counts dataset

Getting residuals for block 12(of 12) for counts dataset

Regressing out percent.rb, percent.mt, CC.Difference, cell_line

Centering and scaling data matrix

Finished calculating residuals for counts

Set default assay to SCT

5. Perform PCA

Variables_genes <- All_samples_Merged@assays$SCT@var.features

# Exclude genes starting with "HLA-" AND "Xist" AND "TRBV, TRAV"
Variables_genes_after_exclusion <- Variables_genes[!grepl("^HLA-|^XIST|^TRBV|^TRAV", Variables_genes)]


# These are now standard steps in the Seurat workflow for visualization and clustering
All_samples_Merged <- RunPCA(All_samples_Merged,
                        features = Variables_genes_after_exclusion,
                        do.print = TRUE, 
                        pcs.print = 1:5, 
                        genes.print = 15,
                        npcs = 50)

PC_ 1 
Positive:  RPS27, CD247, MALAT1, EVL, ETS1, B2M, CD3E, TLE5, PRKCH, RPS29 
       LCK, BCL11B, IL2RG, CTSW, LEF1, NKG7, SKAP1, PRF1, LBH, CCND3 
       CAMK4, TC2N, TBC1D10C, CD3D, RHOH, RORA, TRBC2, LINC00861, TCF7, ABLIM1 
Negative:  S100A9, LYZ, S100A8, TYROBP, VCAN, FCN1, CST3, CTSS, SPI1, CYBB 
       IFI30, PLXDC2, ZEB2, AIF1, FOS, PSAP, HCK, MNDA, LRMDA, NCF2 
       RNF130, RBM47, SERPINA1, RAB31, CD36, CD14, LRP1, ARHGAP26, TBXAS1, CSF3R 
PC_ 2 
Positive:  FCN1, MNDA, CST3, FOS, CSF3R, VCAN, CD36, MS4A6A, CD302, LYZ 
       PRAM1, CLEC12A, KLF4, LRMDA, PLBD1, CFD, CSTA, AOAH, RBP7, FPR1 
       CTSS, ASGR1, AC007952.4, CDA, JAML, TSPO, HK3, CPVL, NFE2, FGR 
Negative:  SOD2, CXCL8, C15orf48, DOCK4, SLC7A11, KYNU, EREG, THBS1, AC025580.2, CXCL5 
       MMP9, SDC2, MMP14, CXCL3, GLIS3, CXCL1, IL1B, SERPINB2, FTH1, CXCL16 
       VMO1, CYP27A1, RAB13, EPB41L3, CTSL, ABCA1, NRP1, IRAK2, NINJ1, CYP1B1 
PC_ 3 
Positive:  CD79A, MS4A1, BANK1, AFF3, IGHM, LINC00926, NIBAN3, CD79B, FCRL1, RALGPS2 
       TCL1A, CD37, CD19, PAX5, HVCN1, FCRLA, CD74, BLNK, GNG7, SPIB 
       IGHD, ADAM28, RUBCNL, LINC02397, CD22, COBLL1, VPREB3, SWAP70, MEF2C, POU2AF1 
Negative:  CD3E, GIMAP7, TCF7, FYB1, GIMAP5, GIMAP4, LEF1, IL7R, GIMAP1, LINC00861 
       SARAF, LCK, IFITM1, TPT1, CD247, ITK, CAMK4, TC2N, CD3D, BCL11B 
       ITM2B, RPS15A, PRKCH, PCED1B-AS1, AAK1, PTPRC, CD3G, IL32, CD2, PRKCQ-AS1 
PC_ 4 
Positive:  PFN1, TUBA1B, RAN, PRELID1, HSPE1, H2AFZ, CHCHD2, ACTB, ATP5MC3, NME1 
       RANBP1, NPM1, PPIA, ATP5F1B, HSPD1, SRM, GAPDH, HSP90AA1, UBE2S, SNRPD1 
       NDUFAB1, PRDX1, CYC1, EIF4A1, TPI1, CYCS, FCER1G, PPP1R14B, ALYREF, NME2 
Negative:  FOXP1, ARHGAP15, MBNL1, ANKRD44, BACH2, RIPOR2, ZBTB20, TNRC6B, RABGAP1L, MAML2 
       MALAT1, PRKCA, ARID1B, BCL2, PDE7A, RUNX1, LRBA, NCOA3, PACS1, ZCCHC7 
       DOCK10, CAMK4, AFF3, ELMO1, FTX, VPS13B, INPP4B, PDE3B, MGAT5, CAMK2D 
PC_ 5 
Positive:  LGALS1, S100A11, S100A4, B2M, S100A6, TMSB4X, IL32, LSP1, SH3BGRL3, TMSB10 
       CRIP1, LAPTM5, CYBA, IFITM2, MYL6, EMP3, TAGLN2, VIM, CD52, IL2RG 
       APOBEC3G, FXYD5, RHOC, TNFRSF18, S1PR4, ACTB, IFITM1, S100A10, ITGB7, LGALS3 
Negative:  NPM1, HSPD1, SRM, HSP90AB1, HSPE1, NME1, HMGA1, HSPA9, RANBP1, HNRNPAB 
       PRKDC, NME2, NCL, HSP90AA1, SERBP1, PRELID1, RAN, CYC1, VDAC1, CCT8 
       UBE2S, TOMM40, RBM17, PPP1R14B, MTDH, PHB, MRPL12, H2AFZ, CCT5, CCT6A 

# determine dimensionality of the data
ElbowPlot(All_samples_Merged, ndims = 50)

6. Perform PCA TEST

library(ggplot2)
library(RColorBrewer)  

# Assuming you have 10 different cell lines, generating a color palette with 10 colors
cell_line_colors <- brewer.pal(10, "Set3")

# Assuming All_samples_Merged$cell_line is a factor or character vector containing cell line names
data <- as.data.frame(table(All_samples_Merged$cell_line))
colnames(data) <- c("cell_line", "nUMI")  # Change column name to nUMI

ncells <- ggplot(data, aes(x = cell_line, y = nUMI, fill = cell_line)) + 
  geom_col() +
  theme_classic() +
  geom_text(aes(label = nUMI), 
            position = position_dodge(width = 0.9), 
            vjust = -0.25) +
  scale_fill_manual(values = cell_line_colors) + 
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        plot.title = element_text(hjust = 0.5)) +  # Adjust the title position
  ggtitle("Filtered cells per sample") +
  xlab("Cell lines") +  # Adjust x-axis label
  ylab("Frequency")    # Adjust y-axis label

print(ncells)

# TEST-1
# given that the output of RunPCA is "pca"
# replace "so" by the name of your seurat object

pct <- All_samples_Merged[["pca"]]@stdev / sum(All_samples_Merged[["pca"]]@stdev) * 100
cumu <- cumsum(pct) # Calculate cumulative percents for each PC
# Determine the difference between variation of PC and subsequent PC
co2 <- sort(which((pct[-length(pct)] - pct[-1]) > 0.1), decreasing = T)[1] + 1
# last point where change of % of variation is more than 0.1%. -> co2
co2

[1] 12

# TEST-2
# get significant PCs
stdv <- All_samples_Merged[["pca"]]@stdev
sum.stdv <- sum(All_samples_Merged[["pca"]]@stdev)
percent.stdv <- (stdv / sum.stdv) * 100
cumulative <- cumsum(percent.stdv)
co1 <- which(cumulative > 90 & percent.stdv < 5)[1]
co2 <- sort(which((percent.stdv[1:length(percent.stdv) - 1] - 
                       percent.stdv[2:length(percent.stdv)]) > 0.1), 
              decreasing = T)[1] + 1
min.pc <- min(co1, co2)
min.pc

[1] 12

# Create a dataframe with values
plot_df <- data.frame(pct = percent.stdv, 
           cumu = cumulative, 
           rank = 1:length(percent.stdv))

# Elbow plot to visualize 
  ggplot(plot_df, aes(cumulative, percent.stdv, label = rank, color = rank > min.pc)) + 
  geom_text() + 
  geom_vline(xintercept = 90, color = "grey") + 
  geom_hline(yintercept = min(percent.stdv[percent.stdv > 5]), color = "grey") +
  theme_bw()

7. Clustering

All_samples_Merged <- FindNeighbors(All_samples_Merged, 
                                dims = 1:12, 
                                verbose = FALSE)

# understanding resolution
All_samples_Merged <- FindClusters(All_samples_Merged, 
                                    resolution = c(0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8, 0.9, 1,1.2,1.5,2))

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.9585
Number of communities: 10
Elapsed time: 48 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.9345
Number of communities: 17
Elapsed time: 43 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.9185
Number of communities: 18
Elapsed time: 44 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.9069
Number of communities: 19
Elapsed time: 38 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8988
Number of communities: 20
Elapsed time: 34 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8894
Number of communities: 21
Elapsed time: 31 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8810
Number of communities: 23
Elapsed time: 25 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8751
Number of communities: 24
Elapsed time: 32 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8668
Number of communities: 29
Elapsed time: 28 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8609
Number of communities: 29
Elapsed time: 21 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8515
Number of communities: 33
Elapsed time: 21 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8387
Number of communities: 35
Elapsed time: 18 seconds
Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1766655

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.8203
Number of communities: 44
Elapsed time: 20 seconds

# non-linear dimensionality reduction --------------
All_samples_Merged <- RunUMAP(All_samples_Merged, 
                          dims = 1:12,
                          verbose = FALSE)

Warning: The default method for RunUMAP has changed from calling Python UMAP via reticulate to the R-native UWOT using the cosine metric
To use Python UMAP via reticulate, set umap.method to 'umap-learn' and metric to 'correlation'
This message will be shown once per session

# note that you can set `label = TRUE` or use the Label Clusters function to help label
# individual clusters
DimPlot(All_samples_Merged,group.by = "cell_line", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,group.by = "predicted.celltype.l2", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.1",
        reduction = "umap",
        label.size = 3,
        repel = T, 
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.2",
        reduction = "umap",
        label.size = 3,
        repel = T, 
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.3",
        reduction = "umap",
        label.size = 3,
        repel = T, 
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.4", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.5", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.6", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.7", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.8", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.0.9", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.1", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.1.2", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.1.5", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged,
        group.by = "SCT_snn_res.2", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

# Set identity classes to an existing column in meta data
Idents(object = All_samples_Merged) <- "SCT_snn_res.0.7"

cluster_table <- table(Idents(All_samples_Merged))


barplot(cluster_table, main = "Number of Cells in Each Cluster", 
                      xlab = "Cluster", 
                      ylab = "Number of Cells", 
                      col = rainbow(length(cluster_table)))

print(cluster_table)

   0    1    2    3    4    5    6    7    8    9   10   11   12   13   14   15 
9572 8683 7501 6356 5637 4673 4671 2335 2219 1378 1150  947  926  794  656  567 
  16   17   18   19   20   21   22 
 395  235  204  178  160   89   29 

table(All_samples_Merged$predicted.celltype.l2, All_samples_Merged$SCT_snn_res.0.2)

                   
                        0     1     2     3     4     5     6     7     8     9
  ASDC                  0     0     0     0     0     0     0     0     0     0
  B intermediate        0     1     0     2     0     0     0   442   174    17
  B memory             34     0     1   207     8     1     2   157    71     4
  B naive               0     0     0     0     0     2     0   471   677     0
  CD14 Mono             2    45    10    13     2  2979     1    14     0   736
  CD16 Mono             1     0     0     0     0   124     0     0     0     1
  CD4 CTL               0    17     0     0     0     0     0     0     0     0
  CD4 Naive             0   523  1476     0     0     1     0     0     0     0
  CD4 Proliferating 23887     0    19  2841   844     0  1256     0     0     0
  CD4 TCM            2397  4481  1866  1404  2335    13    33    29     2    33
  CD4 TEM               0    68    25     1     0     0     0     0     0     0
  CD8 Naive             0   356   995     0     1     1     0     1     1     0
  CD8 Proliferating     1     0     0     1     0     0     0     0     0     0
  CD8 TCM               0   283   173     0    17     0     0     2     0     0
  CD8 TEM               0   180   183     7     5     1     0     1     0     0
  cDC1                  1     0     0     6     0    13     0     0     0     0
  cDC2                  2     0     0    48     1   123     0     1     0     0
  dnT                   2    29    21     7     0     3     0     1     0     2
  gdT                   0    26    67     0     0     0     0     0     0     0
  HSPC               1749    20     1    37    10     0     4     4     0     0
  ILC                   0     1     3     1     0     0     0     1     0     0
  MAIT                  0    15   224     0     0     1     0     0     0     0
  NK                    0    92    16     0     0     5     0     4     0     0
  NK Proliferating   2866     1     4    60   108     0    90     0     0     0
  NK_CD56bright         0     1     5     0     0     0     0     0     0     0
  pDC                   0     0     0     0     0     0     0     0     0     0
  Plasmablast           0     0     0     0     0     0     0    19     0     0
  Platelet              0     3     0     0     0     0     0     0     0     1
  Treg                 10   192   104     3     0     0     0     2     1     0
                   
                       10    11    12    13    14    15    16
  ASDC                  0     0     0     0     0     3     0
  B intermediate        0     2     0    56     2     0     0
  B memory              0     1     0    34     3     0     0
  B naive               0     0     0    42     0     0     0
  CD14 Mono             0     0     0    13     6     2     0
  CD16 Mono             0     0     0     0     0     0     0
  CD4 CTL               0     0     0     0     0     0     0
  CD4 Naive             0     0    36     7     0     0     0
  CD4 Proliferating     0   159     4     1     0     0     0
  CD4 TCM               0    66   176    41     2     0     0
  CD4 TEM               0     0     0     0     0     0     0
  CD8 Naive             0     0    17     0     0     0     1
  CD8 Proliferating     0     0     0     0     0     0     0
  CD8 TCM               0     0     2     0     0     0     0
  CD8 TEM              10     2     2     0     0     0     0
  cDC1                  0     1     0     0    21     0     0
  cDC2                  0     0     0     1    53     0     0
  dnT                   0     5    12     0     0     0     0
  gdT                   0     0     0     0     0     0     0
  HSPC                  0     2     0     6     1     0     0
  ILC                   0     1     0     0     0     0     0
  MAIT                  0     0     2     0     0     0     0
  NK                  414     0     2     1     0     0     0
  NK Proliferating      2    33     3     0     0     0     0
  NK_CD56bright         8     0     2     0     0     0     0
  pDC                   0     0     0     0     0    56     0
  Plasmablast           0     0     0     0     0     0     0
  Platelet              0     0     0     0     0     0    28
  Treg                  0    30     8     2     1     0     0

8. clusTree

clustree(All_samples_Merged, prefix = "SCT_snn_res.")

9. Azimuth Annotation

# InstallData("pbmcref")
# 
# # The RunAzimuth function can take a Seurat object as input
# All_samples_Merged <- RunAzimuth(All_samples_Merged, reference = "pbmcref")

10. Azimuth Visualization

DimPlot(All_samples_Merged, group.by = "predicted.celltype.l1", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged, group.by = "predicted.celltype.l1", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = F)

DimPlot(All_samples_Merged, group.by = "predicted.celltype.l2", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

DimPlot(All_samples_Merged, group.by = "predicted.celltype.l2", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = F)

DimPlot(All_samples_Merged, group.by = "predicted.celltype.l2", 
        reduction = "umap",
        label.size = 3,
        repel = T,
        label = T, label.box = T)

table(All_samples_Merged$predicted.celltype.l2, All_samples_Merged$SCT_snn_res.0.2)

                   
                        0     1     2     3     4     5     6     7     8     9
  ASDC                  0     0     0     0     0     0     0     0     0     0
  B intermediate        0     1     0     2     0     0     0   442   174    17
  B memory             34     0     1   207     8     1     2   157    71     4
  B naive               0     0     0     0     0     2     0   471   677     0
  CD14 Mono             2    45    10    13     2  2979     1    14     0   736
  CD16 Mono             1     0     0     0     0   124     0     0     0     1
  CD4 CTL               0    17     0     0     0     0     0     0     0     0
  CD4 Naive             0   523  1476     0     0     1     0     0     0     0
  CD4 Proliferating 23887     0    19  2841   844     0  1256     0     0     0
  CD4 TCM            2397  4481  1866  1404  2335    13    33    29     2    33
  CD4 TEM               0    68    25     1     0     0     0     0     0     0
  CD8 Naive             0   356   995     0     1     1     0     1     1     0
  CD8 Proliferating     1     0     0     1     0     0     0     0     0     0
  CD8 TCM               0   283   173     0    17     0     0     2     0     0
  CD8 TEM               0   180   183     7     5     1     0     1     0     0
  cDC1                  1     0     0     6     0    13     0     0     0     0
  cDC2                  2     0     0    48     1   123     0     1     0     0
  dnT                   2    29    21     7     0     3     0     1     0     2
  gdT                   0    26    67     0     0     0     0     0     0     0
  HSPC               1749    20     1    37    10     0     4     4     0     0
  ILC                   0     1     3     1     0     0     0     1     0     0
  MAIT                  0    15   224     0     0     1     0     0     0     0
  NK                    0    92    16     0     0     5     0     4     0     0
  NK Proliferating   2866     1     4    60   108     0    90     0     0     0
  NK_CD56bright         0     1     5     0     0     0     0     0     0     0
  pDC                   0     0     0     0     0     0     0     0     0     0
  Plasmablast           0     0     0     0     0     0     0    19     0     0
  Platelet              0     3     0     0     0     0     0     0     0     1
  Treg                 10   192   104     3     0     0     0     2     1     0
                   
                       10    11    12    13    14    15    16
  ASDC                  0     0     0     0     0     3     0
  B intermediate        0     2     0    56     2     0     0
  B memory              0     1     0    34     3     0     0
  B naive               0     0     0    42     0     0     0
  CD14 Mono             0     0     0    13     6     2     0
  CD16 Mono             0     0     0     0     0     0     0
  CD4 CTL               0     0     0     0     0     0     0
  CD4 Naive             0     0    36     7     0     0     0
  CD4 Proliferating     0   159     4     1     0     0     0
  CD4 TCM               0    66   176    41     2     0     0
  CD4 TEM               0     0     0     0     0     0     0
  CD8 Naive             0     0    17     0     0     0     1
  CD8 Proliferating     0     0     0     0     0     0     0
  CD8 TCM               0     0     2     0     0     0     0
  CD8 TEM              10     2     2     0     0     0     0
  cDC1                  0     1     0     0    21     0     0
  cDC2                  0     0     0     1    53     0     0
  dnT                   0     5    12     0     0     0     0
  gdT                   0     0     0     0     0     0     0
  HSPC                  0     2     0     6     1     0     0
  ILC                   0     1     0     0     0     0     0
  MAIT                  0     0     2     0     0     0     0
  NK                  414     0     2     1     0     0     0
  NK Proliferating      2    33     3     0     0     0     0
  NK_CD56bright         8     0     2     0     0     0     0
  pDC                   0     0     0     0     0    56     0
  Plasmablast           0     0     0     0     0     0     0
  Platelet              0     0     0     0     0     0    28
  Treg                  0    30     8     2     1     0     0

11.Harmony Integration

# Load required libraries
library(Seurat)
library(harmony)

Loading required package: Rcpp

library(ggplot2)

# Run Harmony, adjusting for batch effect using "cell_line" or another grouping variable
All_samples_Merged <- RunHarmony(
  object = All_samples_Merged,
  group.by.vars = "cell_line",  # Replace with the metadata column specifying batch or cell line
  dims.use = 1:12  # Use the same dimensions as PCA
)

Transposing data matrix

Initializing state using k-means centroids initialization

Warning: Quick-TRANSfer stage steps exceeded maximum (= 2967750)

Harmony 1/10

Harmony 2/10

Harmony 3/10

Harmony 4/10

Harmony 5/10

Harmony 6/10

Harmony 7/10

Harmony 8/10

Harmony converged after 8 iterations

# Check results in harmony embeddings
harmony_embeddings <- Embeddings(All_samples_Merged, reduction = "harmony")
head(harmony_embeddings)

                        harmony_1 harmony_2  harmony_3 harmony_4 harmony_5
L1_AAACCTGAGGGCTTCC-1 -10.9197271 1.0011300 -1.2319081 -9.454082  3.516301
L1_AAACCTGGTGCAGGTA-1   7.6595107 0.9463873 -2.6776307 -4.739794  5.901444
L1_AAACCTGGTTAAAGTG-1   7.5707677 1.9597971  0.7453221  2.326972 -2.093935
L1_AAACCTGTCAGGTAAA-1  -0.7097921 0.7860023  2.5913303  6.457295 -2.765241
L1_AAACCTGTCCCTGACT-1  -6.0938986 0.6867671 -3.0973393 -7.239550  0.751797
L1_AAACCTGTCCTTCAAT-1   9.1058174 2.0323890 -1.0437973 -5.754146 -3.945509
                        harmony_6  harmony_7  harmony_8  harmony_9  harmony_10
L1_AAACCTGAGGGCTTCC-1  1.51713713  0.7552333  2.5848829 -1.0376164 -0.07275090
L1_AAACCTGGTGCAGGTA-1 -1.42717219 -1.3921612 -1.4959132 -1.2081711  0.61255052
L1_AAACCTGGTTAAAGTG-1 -2.55322630 -2.5568823 -0.9256684  2.3314433 -5.55643819
L1_AAACCTGTCAGGTAAA-1 -2.02404574 -1.4362104 -1.1135829  1.2684311 -1.19682628
L1_AAACCTGTCCCTGACT-1  0.01592589  0.6259760  1.0997142  0.2578556 -0.06975305
L1_AAACCTGTCCTTCAAT-1 -0.20304726 -0.5940280  1.3364864  0.9663652 -1.22088223
                       harmony_11 harmony_12
L1_AAACCTGAGGGCTTCC-1  0.09490475 -0.6359772
L1_AAACCTGGTGCAGGTA-1 -0.04610716  1.1985832
L1_AAACCTGGTTAAAGTG-1  1.65214265 -0.5767076
L1_AAACCTGTCAGGTAAA-1  0.27688001  0.4592023
L1_AAACCTGTCCCTGACT-1 -0.10860366 -1.3188447
L1_AAACCTGTCCTTCAAT-1  2.26618586 -0.3910166

# Run UMAP on Harmony embeddings
All_samples_Merged <- RunUMAP(All_samples_Merged, reduction = "harmony", dims = 1:12)

21:10:57 UMAP embedding parameters a = 0.9922 b = 1.112

21:10:57 Read 59355 rows and found 12 numeric columns

21:10:57 Using Annoy for neighbor search, n_neighbors = 30

21:10:57 Building Annoy index with metric = cosine, n_trees = 50

0%   10   20   30   40   50   60   70   80   90   100%

[----|----|----|----|----|----|----|----|----|----|

**************************************************|
21:11:04 Writing NN index file to temp file /tmp/RtmpNUwYyW/file2721b5a315404
21:11:04 Searching Annoy index using 1 thread, search_k = 3000
21:11:27 Annoy recall = 100%
21:11:29 Commencing smooth kNN distance calibration using 1 thread with target n_neighbors = 30
21:11:34 Initializing from normalized Laplacian + noise (using RSpectra)
21:11:40 Commencing optimization for 200 epochs, with 2454546 positive edges
21:12:53 Optimization finished

# Optionally, find neighbors and clusters (if you plan to do clustering analysis)
All_samples_Merged <- FindNeighbors(All_samples_Merged, reduction = "harmony", dims = 1:12)

Computing nearest neighbor graph
Computing SNN

All_samples_Merged <- FindClusters(All_samples_Merged, resolution = 0.5)  # Adjust resolution as needed

Modularity Optimizer version 1.3.0 by Ludo Waltman and Nees Jan van Eck

Number of nodes: 59355
Number of edges: 1834735

Running Louvain algorithm...
Maximum modularity in 10 random starts: 0.9184
Number of communities: 21
Elapsed time: 27 seconds

# Visualize UMAP
DimPlot(All_samples_Merged, reduction = "umap", group.by = "cell_line", label = TRUE, pt.size = 0.5) + 
    ggtitle("UMAP of Harmony-Integrated Data")

# Visualize UMAP with batch/cell line information
DimPlot(All_samples_Merged, reduction = "umap", group.by = "cell_line", label = TRUE, pt.size = 0.5) + 
    ggtitle("UMAP - Colored by Cell Line (After Harmony Integration)")

# Visualize UMAP with clusters
DimPlot(All_samples_Merged, reduction = "umap", group.by = "seurat_clusters", label = TRUE, pt.size = 0.5) + 
    ggtitle("UMAP - Clustered Data (After Harmony Integration)")

# Visualize specific cell types or other metadata
DimPlot(All_samples_Merged, reduction = "umap", group.by = "predicted.celltype.l2", label = TRUE, pt.size = 0.5) + 
    ggtitle("UMAP - Cell Types After Harmony Integration")

12.Save the Seurat object as an Robj file

save(All_samples_Merged, file = "../../../0-IMP-OBJECTS/All_Samples_Merged_with_10x_Azitmuth_Annotated_SCT_HPC_SCT-regressed-cell_line-1-12.robj")