UMAP_visualization

1 load libraries

2 Load Seurat Object

All_samples_Merged <- readRDS("../../../PHD_3rd_YEAR_Analysis/0-Seurat_RDS_OBJECT_FINAL/All_samples_Merged_with_Renamed_Clusters_Cell_state-03-12-2025.rds.rds")

ss <- All_samples_Merged

rm(All_samples_Merged)

gc()

             used   (Mb) gc trigger    (Mb)   max used   (Mb)
Ncells    9806904  523.8   14314937   764.6   14314937  764.6
Vcells 1234922613 9421.8 1533268137 11698.0 1235214144 9424.0

ss

An object of class Seurat 
62900 features across 49305 samples within 6 assays 
Active assay: RNA (36601 features, 0 variable features)
 2 layers present: data, counts
 5 other assays present: ADT, prediction.score.celltype.l1, prediction.score.celltype.l2, prediction.score.celltype.l3, SCT
 5 dimensional reductions calculated: integrated_dr, ref.umap, pca, umap, harmony

3 Create the base UMAP using SCpubr with ALL original populations

# =============================================================================
# SIMPLIFIED APPROACH: Label only cell line regions (L1/L2, L3/L4, L5/L6/L7)
# with P1, P2, P3, and normal region with "Normal CD4 T cells"
# =============================================================================

# Step 1: Base plot
p <- SCpubr::do_DimPlot(
  sample = seurat_obj,
  group.by = "orig.ident",
  label = FALSE,
  legend.position = "right",
   plot.axes = T,
  pt.size = 0.8,
  raster = FALSE
)

# Step 2: UMAP coords (lowercase columns)
umap_coords <- as.data.frame(seurat_obj@reductions$umap@cell.embeddings)
umap_coords$orig.ident <- seurat_obj$orig.ident

# Step 3: Calculate centroids
label_positions <- data.frame(label = character(), umap_1 = numeric(), umap_2 = numeric())

# P1: L1+L2
p1_cells <- umap_coords[umap_coords$orig.ident %in% c("L1", "L2"), ]
if(nrow(p1_cells) > 0) label_positions <- rbind(label_positions,
  data.frame(label = "P1", 
             umap_1 = median(p1_cells$umap_1, na.rm = TRUE),
             umap_2 = median(p1_cells$umap_2, na.rm = TRUE)))

# P2: L3+L4
p2_cells <- umap_coords[umap_coords$orig.ident %in% c("L3", "L4"), ]
if(nrow(p2_cells) > 0) label_positions <- rbind(label_positions,
  data.frame(label = "P2", 
             umap_1 = median(p2_cells$umap_1, na.rm = TRUE),
             umap_2 = median(p2_cells$umap_2, na.rm = TRUE)))

# P3: L5+L6+L7
p3_cells <- umap_coords[umap_coords$orig.ident %in% c("L5", "L6", "L7"), ]
if(nrow(p3_cells) > 0) label_positions <- rbind(label_positions,
  data.frame(label = "P3", 
             umap_1 = median(p3_cells$umap_1, na.rm = TRUE),
             umap_2 = median(p3_cells$umap_2, na.rm = TRUE)))

# Normal CD4 T cells
normal_cells <- umap_coords[umap_coords$orig.ident %in% c("CD4T_lab", "CD4T_10x"), ]
if(nrow(normal_cells) > 0) label_positions <- rbind(label_positions,
  data.frame(label = "Normal CD4 T cells", 
             umap_1 = median(normal_cells$umap_1, na.rm = TRUE),
             umap_2 = median(normal_cells$umap_2, na.rm = TRUE)))

print("Label positions:")

[1] "Label positions:"

print(label_positions)

# Step 4: Final plot with UMAP1/UMAP2 axes
p_final <- p + 
  geom_text_repel(
    data = label_positions,
    aes(x = umap_1, y = umap_2, label = label),
    inherit.aes = FALSE,
    size = 6, fontface = "bold", color = "black",
    bg.color = "white", bg.r = 0.15,
    box.padding = 0.5, point.padding = 0.5,
    min.segment.length = 0, max.overlaps = Inf
  ) +
  xlab("UMAP_1") + 
  ylab("UMAP_2") +
  theme(axis.title = element_text(size = 14, face = "bold"))

# Display
print(p_final)

# Step 5: Save BOTH PDF and PNG
ggsave("UMAP_P1_P2_P3_Normal.pdf", p_final, width = 12, height = 8, dpi = 300)
ggsave("UMAP_P1_P2_P3_Normal.png", p_final, width = 12, height = 8, dpi = 300)

cat("\n✅ Saved both files:\n")

✅ Saved both files:

cat("   📄 UMAP_P1_P2_P3_Normal.pdf\n")

   📄 UMAP_P1_P2_P3_Normal.pdf

cat("   🖼️  UMAP_P1_P2_P3_Normal.png\n")

   🖼️  UMAP_P1_P2_P3_Normal.png

4 Create the base UMAP using SCpubr with ALL original populations

5 Create the base UMAP using SCpubr with ALL original populations

# =============================================================================
# SIMPLIFIED: Show only cluster numbers, no P1/P2/P3 labels
# =============================================================================

# Step 1: Base plot with cluster labels
p_cluster <- SCpubr::do_DimPlot(
  sample = seurat_obj,
  group.by = "seurat_clusters",
  label = TRUE,
  legend.position = "right",
  label.color = "white",
  plot.axes = TRUE,
  pt.size = 0.8,
  raster = FALSE
)

# Step 2: Add axis labels (no region annotations)
p2 <- p_cluster + 
  xlab("UMAP_1") + 
  ylab("UMAP_2") +
  theme(axis.title = element_text(size = 14, face = "bold"))

# Step 3: Display
print(p2)

# Step 4: Save both PDF and PNG
ggsave("UMAP_Clusters_only.pdf", p2, width = 12, height = 8, dpi = 300)
ggsave("UMAP_Clusters_only.png", p2, width = 12, height = 8, dpi = 300)

cat("\n✅ Saved both files:\n")

✅ Saved both files:

cat("   📄 UMAP_Clusters_only.pdf\n")

   📄 UMAP_Clusters_only.pdf

cat("   🖼️  UMAP_Clusters_only.png\n")

   🖼️  UMAP_Clusters_only.png

6 Th Polarization

genes <- list(
  Th1 = c("TBX21", "CXCR3", "CXCR6", "IFNG"),
  Th2 = c("GATA3", "CCR4", "PTGDR2", "IL4"),
  Th17_Th22 = c("RORC", "CCR6", "CCR10", "RORA", "STAT3", "IL17A", "IL21"),
  Treg = c("FOXP3", "IL2RA", "CTLA4")
)

p2 <- SCpubr::do_DotPlot(
  sample = seurat_obj, group.by = "orig.ident",
  features = genes
)

p2

7 Th Polarization-flip

genes <- list(
  Th1 = c("TBX21", "CXCR3", "CXCR6", "IFNG"),
  Th2 = c("GATA3", "CCR4", "PTGDR2", "IL4"),
  Th17_Th22 = c("RORC", "CCR6", "CCR10", "RORA", "STAT3", "IL17A", "IL21"),
  Treg = c("FOXP3", "IL2RA", "CTLA4")
)


p3 <- SCpubr::do_DotPlot(sample = seurat_obj, group.by = "orig.ident",flip = T,
                        features = genes)

p3