---
title: "Visualization Gallery"
author: "Zaoqu Liu, Robert K. Suter, Nagi G. Ayad"
date: "`r Sys.Date()`"
output: 
  rmarkdown::html_vignette:
    toc: true
    toc_depth: 3
vignette: >
  %\VignetteIndexEntry{Visualization Gallery}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(
  echo = TRUE,
  message = FALSE,
  warning = FALSE,
  fig.width = 8,
  fig.height = 6,
  eval = TRUE
)
library(ggplot2)
library(dplyr)
library(tidyr)
library(viridis)
library(pheatmap)
```

## Introduction

scFOCAL generates a variety of publication-ready visualizations. This gallery showcases the main output types with customization examples.

## 1. Dimensional Reduction Plots

### UMAP with Cell Type Annotations

```{r umap-celltype, fig.height=5, fig.width=7}
set.seed(42)
n_cells <- 1500

# Generate mock UMAP data
umap_data <- data.frame(
  UMAP1 = c(rnorm(300, -3, 0.8), rnorm(300, 0, 0.9), rnorm(300, 3, 0.8),
            rnorm(300, -1.5, 0.7), rnorm(300, 1.5, 0.7)),
  UMAP2 = c(rnorm(300, 2, 0.7), rnorm(300, 3, 0.8), rnorm(300, 1, 0.7),
            rnorm(300, -2, 0.8), rnorm(300, -1, 0.7)),
  CellType = factor(rep(c("Mesenchymal", "Astrocyte-like", "Neural Progenitor",
                          "Immune", "Stromal"), each = 300))
)

ggplot(umap_data, aes(x = UMAP1, y = UMAP2, color = CellType)) +
  geom_point(size = 0.8, alpha = 0.7) +
  scale_color_brewer(palette = "Set2", name = "Cell Type") +
  labs(
    title = "Single-Cell UMAP Projection",
    subtitle = "Cell type annotations",
    x = "UMAP 1",
    y = "UMAP 2"
  ) +
  theme_minimal() +
  theme(
    legend.position = "right",
    plot.title = element_text(face = "bold", size = 14)
  ) +
  guides(color = guide_legend(override.aes = list(size = 3)))
```

### UMAP with Drug Connectivity

```{r umap-connectivity, fig.height=5, fig.width=8}
umap_data$Connectivity <- c(
  rnorm(300, -0.15, 0.1),   # MES - sensitive
  rnorm(300, 0.1, 0.12),    # AC - resistant
  rnorm(300, 0.05, 0.1),    # NPC
  rnorm(300, 0, 0.08),      # Immune
  rnorm(300, 0, 0.08)       # Stromal
)

ggplot(umap_data, aes(x = UMAP1, y = UMAP2, color = Connectivity)) +
  geom_point(size = 0.8, alpha = 0.7) +
  scale_color_gradient2(
    low = "#2166AC",
    mid = "white",
    high = "#B2182B",
    midpoint = 0,
    name = "Drug\nConnectivity"
  ) +
  labs(
    title = "Drug-Cell Connectivity Score",
    subtitle = "Temozolomide response prediction",
    x = "UMAP 1",
    y = "UMAP 2"
  ) +
  theme_minimal() +
  theme(plot.title = element_text(face = "bold", size = 14))
```

## 2. Violin Plots

### Drug Connectivity by Cell Type

```{r violin-basic, fig.height=5, fig.width=7}
tumor_data <- umap_data[umap_data$CellType %in% c("Mesenchymal", "Astrocyte-like", "Neural Progenitor"),]

ggplot(tumor_data, aes(x = CellType, y = Connectivity, fill = CellType)) +
  geom_violin(alpha = 0.7, scale = "width") +
  geom_boxplot(width = 0.15, fill = "white", outlier.size = 0.5) +
  geom_hline(yintercept = 0, linetype = "dashed", color = "gray40") +
  scale_fill_brewer(palette = "Set2") +
  labs(
    title = "Drug Sensitivity by Tumor State",
    subtitle = "Violin plot with embedded boxplot",
    x = "",
    y = "Drug-Cell Connectivity"
  ) +
  theme_minimal() +
  theme(
    legend.position = "none",
    axis.text.x = element_text(angle = 30, hjust = 1),
    plot.title = element_text(face = "bold", size = 14)
  ) +
  annotate("text", x = 3.4, y = -0.3, label = "Sensitive", 
           color = "#2166AC", fontface = "italic", size = 3.5) +
  annotate("text", x = 3.4, y = 0.3, label = "Resistant", 
           color = "#B2182B", fontface = "italic", size = 3.5)
```

## 3. Heatmaps

### Disease Signature Heatmap

```{r heatmap-disease-sig, fig.height=6, fig.width=8}
set.seed(123)
# Generate mock expression data
n_genes <- 30
n_cells <- 100

gene_names <- paste0("Gene_", 1:n_genes)
cell_types <- rep(c("MES", "AC", "NPC"), c(35, 35, 30))

# Create expression matrix with cell type-specific patterns
expr_matrix <- matrix(rnorm(n_genes * n_cells), nrow = n_genes, ncol = n_cells)
rownames(expr_matrix) <- gene_names
colnames(expr_matrix) <- paste0("Cell_", 1:n_cells)

# Add cell type-specific patterns
for (i in 1:10) {
  expr_matrix[i, cell_types == "MES"] <- expr_matrix[i, cell_types == "MES"] + 1.5
}
for (i in 11:20) {
  expr_matrix[i, cell_types == "AC"] <- expr_matrix[i, cell_types == "AC"] + 1.5
}
for (i in 21:30) {
  expr_matrix[i, cell_types == "NPC"] <- expr_matrix[i, cell_types == "NPC"] + 1.5
}

# Annotation
annotation_col <- data.frame(
  CellType = factor(cell_types),
  row.names = colnames(expr_matrix)
)

annotation_colors <- list(
  CellType = c(MES = "#66C2A5", AC = "#FC8D62", NPC = "#8DA0CB")
)

pheatmap(
  expr_matrix,
  scale = "row",
  show_colnames = FALSE,
  annotation_col = annotation_col,
  annotation_colors = annotation_colors,
  color = colorRampPalette(c("#2166AC", "white", "#B2182B"))(100),
  main = "Disease Signature Gene Expression",
  fontsize = 10,
  cluster_cols = FALSE
)
```

## 4. Volcano Plots

### Differential Connectivity Volcano

```{r volcano, fig.height=6, fig.width=8}
set.seed(789)
n_compounds <- 1679

compound_names <- c("Temozolomide", "Vincristine", "Doxorubicin", "Erlotinib", 
                    "Dasatinib", "Imatinib", "Sorafenib", "Sunitinib")

volcano_data <- data.frame(
  Compound = paste0("Compound_", 1:n_compounds),
  logFC = rnorm(n_compounds, 0, 0.12),
  P.Value = 10^(-runif(n_compounds, 0, 3))
)

# Add significant hits
sig_up <- sample(1:n_compounds, 30)
sig_down <- sample(setdiff(1:n_compounds, sig_up), 30)
volcano_data$logFC[sig_up] <- runif(30, 0.3, 0.6)
volcano_data$logFC[sig_down] <- runif(30, -0.6, -0.3)
volcano_data$P.Value[c(sig_up, sig_down)] <- 10^(-runif(60, 3, 6))

volcano_data$adj.P.Val <- p.adjust(volcano_data$P.Value, method = "fdr")

volcano_data$Significance <- case_when(
  volcano_data$logFC > 0.2 & volcano_data$adj.P.Val < 0.05 ~ "Up in Resistant",
  volcano_data$logFC < -0.2 & volcano_data$adj.P.Val < 0.05 ~ "Up in Sensitive",
  TRUE ~ "Not Significant"
)

# Add compound labels for top hits
volcano_data$Label <- ""
top_sig <- which(volcano_data$adj.P.Val < 0.001)
volcano_data$Label[top_sig[1:min(8, length(top_sig))]] <- compound_names[1:min(8, length(top_sig))]

ggplot(volcano_data, aes(x = logFC, y = -log10(adj.P.Val))) +
  geom_point(aes(color = Significance), alpha = 0.6, size = 1.5) +
  geom_hline(yintercept = -log10(0.05), linetype = "dashed", color = "gray50") +
  geom_vline(xintercept = c(-0.2, 0.2), linetype = "dashed", color = "gray50") +
  scale_color_manual(values = c(
    "Up in Resistant" = "#B2182B",
    "Up in Sensitive" = "#2166AC",
    "Not Significant" = "gray70"
  )) +
  labs(
    title = "Differential Connectivity Analysis",
    subtitle = "Resistant vs Sensitive cell populations",
    x = "log2 Fold Change",
    y = "-log10(adjusted P-value)"
  ) +
  theme_minimal() +
  theme(
    legend.position = "top",
    plot.title = element_text(face = "bold", size = 14)
  )
```

## 5. Reversal Score Visualization

### Bar Plot of Top Reversal Compounds

```{r reversal-bar, fig.height=5, fig.width=7}
reversal_data <- data.frame(
  Compound = compound_names[1:8],
  Reversal_Score = c(2.8, 2.5, 2.3, 2.1, 1.9, 1.8, 1.7, 1.6),
  MOA = c("Kinase Inhibitor", "DNA Damage", "Epigenetic", "Metabolic", 
          "Kinase Inhibitor", "DNA Damage", "Epigenetic", "Other")
)

ggplot(reversal_data, aes(x = reorder(Compound, Reversal_Score), y = Reversal_Score, fill = MOA)) +
  geom_col(width = 0.7) +
  geom_hline(yintercept = 1, linetype = "dashed", color = "gray40") +
  coord_flip() +
  scale_fill_brewer(palette = "Set2", name = "Mechanism\nof Action") +
  labs(
    title = "Disease Signature Reversal Scores",
    subtitle = "Top compounds for Mesenchymal state reversal",
    x = "",
    y = "Reversal Score (Discordant / Concordant)"
  ) +
  theme_minimal() +
  theme(
    plot.title = element_text(face = "bold", size = 14),
    legend.position = "right"
  ) +
  annotate("text", x = 2, y = 1.5, label = "Reversal threshold", 
           color = "gray40", fontface = "italic", size = 3)
```

## 6. Multi-Panel Publication Figure

```{r composite, fig.height=10, fig.width=12}
# Panel A: UMAP
p1 <- ggplot(umap_data, aes(x = UMAP1, y = UMAP2, color = Connectivity)) +
  geom_point(size = 0.5, alpha = 0.6) +
  scale_color_gradient2(low = "#2166AC", mid = "white", high = "#B2182B", midpoint = 0) +
  labs(title = "A) Drug Connectivity", x = "UMAP 1", y = "UMAP 2") +
  theme_minimal() +
  theme(legend.position = "bottom", plot.title = element_text(face = "bold"))

# Panel B: Violin
p2 <- ggplot(tumor_data, aes(x = CellType, y = Connectivity, fill = CellType)) +
  geom_violin(alpha = 0.7) +
  geom_boxplot(width = 0.1, fill = "white") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  scale_fill_brewer(palette = "Set2") +
  labs(title = "B) By Cell Type", x = "", y = "Connectivity") +
  theme_minimal() +
  theme(legend.position = "none", 
        axis.text.x = element_text(angle = 30, hjust = 1),
        plot.title = element_text(face = "bold"))

# Panel C: Volcano (subset)
p3 <- ggplot(volcano_data, aes(x = logFC, y = -log10(adj.P.Val), color = Significance)) +
  geom_point(alpha = 0.5, size = 1) +
  geom_hline(yintercept = -log10(0.05), linetype = "dashed") +
  geom_vline(xintercept = c(-0.2, 0.2), linetype = "dashed") +
  scale_color_manual(values = c("Up in Resistant" = "#B2182B", 
                                 "Up in Sensitive" = "#2166AC", 
                                 "Not Significant" = "gray70")) +
  labs(title = "C) Differential Connectivity", x = "logFC", y = "-log10(adj.P)") +
  theme_minimal() +
  theme(legend.position = "bottom", plot.title = element_text(face = "bold"))

# Panel D: Reversal
p4 <- ggplot(reversal_data, aes(x = reorder(Compound, Reversal_Score), y = Reversal_Score)) +
  geom_col(fill = "steelblue", width = 0.7) +
  geom_hline(yintercept = 1, linetype = "dashed") +
  coord_flip() +
  labs(title = "D) Top Reversal Compounds", x = "", y = "Reversal Score") +
  theme_minimal() +
  theme(plot.title = element_text(face = "bold"))

gridExtra::grid.arrange(p1, p2, p3, p4, ncol = 2, nrow = 2)
```

## Customization Tips

### Color Palettes

For publication-quality figures, consider these color schemes:

- **Diverging** (connectivity): Blue-White-Red for negative/neutral/positive
- **Sequential** (expression): Viridis for continuous values  
- **Categorical** (cell types): Set2 or Paired from RColorBrewer

### Export Settings

```{r export-settings, eval=FALSE}
# PNG (for slides)
ggsave("figure.png", width = 8, height = 6, dpi = 300)

# PDF (for manuscripts)
ggsave("figure.pdf", width = 8, height = 6)

# TIFF (some journals require)
ggsave("figure.tiff", width = 8, height = 6, dpi = 300, compression = "lzw")
```

## Session Info

```{r session-info}
sessionInfo()
```