Complex UpSet Plot Visualization with jcomplexupset

Introduction to Complex UpSet Plots

Complex UpSet plots are a powerful visualization technique for analyzing intersections among multiple sets. They are particularly useful when you have more than 3 sets, where traditional Venn diagrams become cluttered and difficult to interpret. The jcomplexupset function in ClinicoPath provides an enhanced interface to create sophisticated UpSet plots for clinical and pathological research.

When to Use UpSet Plots

UpSet plots are ideal for:

Treatment Combinations: Analyzing which treatments are used together
Biomarker Patterns: Understanding co-occurrence of multiple biomarkers
Complication Analysis: Identifying common combinations of side effects
Diagnostic Workflows: Visualizing which tests are performed together
Molecular Subtyping: Analyzing mutation patterns in cancer research

Advantages over Venn Diagrams

Scalability: Handle 10+ sets effectively
Quantitative: Show exact intersection sizes
Readable: Matrix layout prevents overlapping regions
Flexible: Support for weighted intersections and complex annotations

Getting Started

Loading Required Packages

library(ClinicoPath)
library(dplyr)
library(ggplot2)

# Load test datasets
data("jcomplexupset_test_data")
data("molecular_subtype_data") 
data("diagnostic_test_data")

Basic Dataset Structure

For UpSet plots, you need:

Binary Variables: Columns representing set membership (TRUE/FALSE)
Sufficient Sample Size: Adequate observations for meaningful intersections
Multiple Sets: At least 2 sets, ideally 3-8 for best visualization

# Examine the treatment data structure
str(jcomplexupset_test_data[, c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy", "TargetedTherapy")])

## 'data.frame':    200 obs. of  5 variables:
##  $ Surgery        : logi  TRUE TRUE TRUE FALSE TRUE TRUE ...
##  $ Chemotherapy   : logi  TRUE TRUE FALSE TRUE TRUE FALSE ...
##  $ Radiotherapy   : logi  TRUE TRUE FALSE TRUE FALSE TRUE ...
##  $ Immunotherapy  : logi  FALSE FALSE FALSE FALSE FALSE FALSE ...
##  $ TargetedTherapy: logi  TRUE FALSE FALSE TRUE FALSE FALSE ...

# Summary of treatment combinations
treatment_vars <- c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy", "TargetedTherapy")
summary(jcomplexupset_test_data[, treatment_vars])

##   Surgery        Chemotherapy    Radiotherapy    Immunotherapy  
##  Mode :logical   Mode :logical   Mode :logical   Mode :logical  
##  FALSE:60        FALSE:78        FALSE:93        FALSE:143      
##  TRUE :140       TRUE :122       TRUE :107       TRUE :57       
##  TargetedTherapy
##  Mode :logical  
##  FALSE:159      
##  TRUE :41

Basic UpSet Plots

Simple Treatment Analysis

Let’s start with a basic analysis of treatment combinations:

# Basic UpSet plot for treatment modalities
# Note: This demonstrates jcomplexupset functionality using ggplot2
# In jamovi, you would use the jcomplexupset analysis from the menu

# Create a simplified demonstration plot
library(ggplot2)

# For demonstration, let's show how treatment combinations would be visualized
treatment_combinations <- jcomplexupset_test_data %>%
  select(Surgery, Chemotherapy, Radiotherapy, Immunotherapy, TargetedTherapy) %>%
  mutate(
    combination = paste(
      ifelse(Surgery == 1, "Surgery", ""),
      ifelse(Chemotherapy == 1, "Chemo", ""),
      ifelse(Radiotherapy == 1, "Radio", ""), 
      ifelse(Immunotherapy == 1, "Immuno", ""),
      ifelse(TargetedTherapy == 1, "Targeted", ""),
      sep = "+"
    )
  ) %>%
  # Clean up the combination strings
  mutate(combination = gsub("\\++", "+", combination)) %>%
  mutate(combination = gsub("^\\+|\\+$", "", combination)) %>%
  filter(combination != "") %>%
  count(combination, sort = TRUE) %>%
  top_n(10)

## Selecting by n

ggplot(treatment_combinations, aes(x = reorder(combination, n), y = n)) +
  geom_col(fill = "#2E86AB", alpha = 0.8) +
  coord_flip() +
  theme_minimal() +
  labs(title = "Treatment Combination Analysis",
       subtitle = "Frequency of different treatment modalities and their combinations",
       x = "Treatment Combination", y = "Frequency") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

Biomarker Pattern Analysis

Analyze biomarker co-occurrence patterns:

# Note: Due to ComplexUpset package dependencies, we'll demonstrate with a simpler visualization
# In practice, jcomplexupset would create an UpSet plot for biomarker patterns

# Create a demonstration of biomarker co-occurrence
biomarker_data <- jcomplexupset_test_data %>%
  select(HER2_Positive, ER_Positive, PR_Positive, PDL1_Positive) %>%
  mutate(
    combination = paste(
      ifelse(HER2_Positive == 1, "HER2+", ""),
      ifelse(ER_Positive == 1, "ER+", ""),
      ifelse(PR_Positive == 1, "PR+", ""), 
      ifelse(PDL1_Positive == 1, "PDL1+", ""),
      sep = "/"
    )
  ) %>%
  mutate(combination = gsub("/+", "/", combination)) %>%
  mutate(combination = gsub("^/|/$", "", combination)) %>%
  filter(combination != "") %>%
  count(combination, sort = TRUE) %>%
  top_n(10)

## Selecting by n

ggplot(biomarker_data, aes(x = reorder(combination, n), y = n)) +
  geom_col(fill = "#E74C3C", alpha = 0.8) +
  coord_flip() +
  theme_minimal() +
  labs(
    title = "Biomarker Expression Patterns",
    subtitle = "Co-occurrence of key biomarkers in cancer patients",
    x = "Biomarker Combination", 
    y = "Frequency"
  ) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

# Note: The jcomplexupset function would normally create an UpSet plot like this:
# jcomplexupset(
#   data = jcomplexupset_test_data,
#   set_vars = c("HER2_Positive", "ER_Positive", "PR_Positive", "PDL1_Positive"),
#   plot_title = "Biomarker Expression Patterns",
#   plot_subtitle = "Co-occurrence of key biomarkers in cancer patients"
# )

Advanced Visualization Options

Custom Themes and Colors

# Note: Professional publication-ready styling with jcomplexupset
# Due to ComplexUpset package issues, showing alternative visualization

# Create treatment combination analysis
treatment_styled <- jcomplexupset_test_data %>%
  select(Surgery, Chemotherapy, Radiotherapy, Immunotherapy) %>%
  mutate(
    combination = paste(
      ifelse(Surgery == 1, "Surgery", ""),
      ifelse(Chemotherapy == 1, "Chemo", ""),
      ifelse(Radiotherapy == 1, "Radio", ""), 
      ifelse(Immunotherapy == 1, "Immuno", ""),
      sep = "+"
    )
  ) %>%
  mutate(combination = gsub("\\++", "+", combination)) %>%
  mutate(combination = gsub("^\\+|\\+$", "", combination)) %>%
  filter(combination != "", rowSums(select(., Surgery:Immunotherapy)) >= 1) %>%
  count(combination, sort = TRUE) %>%
  filter(n >= 3) %>%
  top_n(10)

## Selecting by n

ggplot(treatment_styled, aes(x = reorder(combination, n), y = n)) +
  geom_col(fill = "#440154", alpha = 0.8) +
  coord_flip() +
  theme_minimal() +
  labs(
    title = "Treatment Combinations in Oncology",
    subtitle = "Analysis of multimodal treatment approaches",
    x = "Treatment Combination", 
    y = "Frequency"
  ) +
  theme(
    text = element_text(size = 12),
    axis.text = element_text(size = 11),
    plot.title = element_text(size = 14, face = "bold")
  )

# Note: The actual jcomplexupset call would be:
# jcomplexupset(
#   data = jcomplexupset_test_data,
#   set_vars = c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy"),
#   theme_style = "theme_minimal",
#   color_palette = "viridis",
#   text_size = 12,
#   min_size = 3,
#   plot_title = "Treatment Combinations in Oncology",
#   plot_subtitle = "Analysis of multimodal treatment approaches"
# )

Adding Annotations

# Note: UpSet plot with intersection size annotations would normally use jcomplexupset
# Due to ComplexUpset dependency issues, showing summary instead

# Analysis of treatment response patterns
response_vars <- c("Complete_Response", "Partial_Response", "Stable_Disease", "Progressive_Disease")
response_summary <- jcomplexupset_test_data[, response_vars]

# Check response patterns (should be mutually exclusive)
response_counts <- colSums(response_summary, na.rm = TRUE)
cat("Treatment Response Distribution:\n")

## Treatment Response Distribution:

print(response_counts)

##   Complete_Response    Partial_Response      Stable_Disease Progressive_Disease 
##                  39                  46                  29                  23

# Create visualization of response distribution
response_df <- data.frame(
  Response = names(response_counts),
  Count = as.numeric(response_counts)
)

ggplot(response_df, aes(x = reorder(Response, Count), y = Count)) +
  geom_col(fill = c("#E8F5E8", "#FFEB3B", "#FF9800", "#F44336"), alpha = 0.8) +
  coord_flip() +
  theme_classic() +
  labs(
    title = "Treatment Response Patterns",
    subtitle = "Distribution of response categories (mutually exclusive)",
    x = "Response Category", 
    y = "Number of Patients"
  ) +
  theme(
    plot.title = element_text(size = 14, face = "bold"),
    axis.text = element_text(size = 11)
  )

# Note: Actual jcomplexupset call would be:
# jcomplexupset(
#   data = jcomplexupset_test_data,
#   set_vars = c("Complete_Response", "Partial_Response", "Stable_Disease", "Progressive_Disease"),
#   annotations = "intersection_size",
#   base_annotations_height = 2,
#   min_size = 1,
#   theme_style = "theme_classic",
#   color_palette = "Set2",
#   plot_title = "Treatment Response Patterns",
#   plot_subtitle = "Distribution of response categories (mutually exclusive)"
# )

Clinical Applications

Molecular Subtyping Analysis

Analyze mutation patterns in molecular subtyping:

# Mutation pattern analysis
mutation_vars <- c("BRCA1_Mutation", "BRCA2_Mutation", "TP53_Mutation", "PIK3CA_Mutation", "KRAS_Mutation")

# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = molecular_subtype_data,
#   set_vars = mutation_vars,
#   value_var = NULL,
#   min_size = 2,
#   sort_by = "cardinality",
#   show_percentages = FALSE,
#   theme_style = "theme_bw",
#   color_palette = "Dark2",
#   plot_title = "Genetic Mutation Patterns",
#   plot_subtitle = "Co-occurrence of driver mutations in cancer samples"
# )

Pathway Analysis

# Pathway alteration patterns
pathway_vars <- c("PI3K_Pathway", "WNT_Pathway", "RB_Pathway", "DNA_Repair")

# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = molecular_subtype_data,
#   set_vars = pathway_vars,
#   value_var = NULL,
#   annotations = "intersection_ratio",
#   sort_by = "degree",
#   keep_empty_groups = FALSE,
#   plot_title = "Pathway Alteration Patterns",
#   plot_subtitle = "Simultaneous pathway disruptions in cancer"
# )

Diagnostic Workflow Analysis

Analyze patterns in diagnostic test utilization:

# Imaging modality combinations
imaging_vars <- c("CT_Scan", "MRI", "PET_Scan", "Ultrasound", "Biopsy")

# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = diagnostic_test_data,
#   set_vars = imaging_vars,
#   value_var = NULL,
#   min_size = 10,
#   max_degree = 4,
#   sort_by = "cardinality",
#   theme_style = "theme_minimal",
#   color_palette = "plasma",
#   plot_title = "Diagnostic Imaging Combinations",
#   plot_subtitle = "Patterns of imaging modality utilization"
# )

Advanced Features

Statistical Analysis Integration

# UpSet plot with comprehensive statistics
# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = jcomplexupset_test_data,
#   set_vars = c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy"),
#   value_var = NULL,
#   show_statistics = TRUE,
#   show_interpretation = TRUE,
#   min_size = 5,
#   annotations = "intersection_size",
#   plot_title = "Treatment Analysis with Statistics",
#   plot_subtitle = "Comprehensive analysis of treatment combinations"
# )

Customizing Plot Dimensions

# Adjust plot proportions for better visualization
# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = diagnostic_test_data,
#   set_vars = c("Oncology", "Surgery", "Radiology", "Pathology"),
#   value_var = NULL,
#   width_ratio = 0.4,
#   height_ratio = 0.6,
#   sort_by = "cardinality",
#   theme_style = "theme_classic",
#   plot_title = "Specialist Consultation Patterns",
#   plot_subtitle = "Multi-disciplinary care coordination"
# )

Interpretation Guidelines

Reading UpSet Plots

Set Sizes (Left): Individual bars show how many elements are in each set
Intersection Matrix (Bottom): Dots connected by lines show which sets are being intersected
Intersection Sizes (Top): Bars show the size of each intersection
Sorting: Intersections are typically sorted by size (largest first) or degree (number of sets)

Clinical Insights

Treatment Combinations

Monotherapy vs. Combination: Compare single treatments vs. multi-modal approaches
Standard vs. Novel: Identify emerging treatment patterns
Resource Utilization: Understand treatment complexity and cost implications

Biomarker Patterns

Molecular Subtypes: Identify distinct molecular patterns
Prognostic Groups: Correlate patterns with outcomes
Treatment Selection: Guide precision medicine approaches

Diagnostic Workflows

Efficiency Analysis: Identify redundant or missing tests
Quality Indicators: Assess completeness of workups
Cost Optimization: Streamline diagnostic pathways

Best Practices

Data Preparation

# Example of data preprocessing for UpSet analysis
treatment_data <- jcomplexupset_test_data %>%
  # Ensure binary variables are logical
  mutate(
    across(c(Surgery, Chemotherapy, Radiotherapy, Immunotherapy), as.logical)
  ) %>%
  # Remove cases with no treatments (if needed)
  filter(
    Surgery | Chemotherapy | Radiotherapy | Immunotherapy
  )

# Check data quality
cat("Sample size after preprocessing:", nrow(treatment_data), "\n")

## Sample size after preprocessing: 193

cat("Treatment coverage:\n")

## Treatment coverage:

print(colSums(treatment_data[, c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy")]))

##       Surgery  Chemotherapy  Radiotherapy Immunotherapy 
##           140           122           107            57

Visualization Guidelines

Set Selection: Use 3-8 sets for optimal readability
Minimum Size: Filter small intersections (min_size ≥ 3-5)
Sorting: Choose appropriate sorting (size vs. degree vs. ratio)
Colors: Use colorblind-friendly palettes
Annotations: Add relevant statistical information

Parameter Optimization

# Optimized parameters for clinical data
# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = jcomplexupset_test_data,
#   set_vars = c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy", "TargetedTherapy"),
#   value_var = NULL,
#   min_size = 8,           # Filter small intersections
#   max_degree = 3,         # Limit complexity
#   sort_by = "cardinality",       # Most interpretable sorting
#   sort_order = "descending",
#   keep_empty_groups = FALSE,  # Remove empty intersections
#   show_percentages = FALSE,    # Add percentage information
#   theme_style = "theme_minimal",
#   color_palette = "viridis",
#   text_size = 11,
#   plot_title = "Optimized Treatment Combination Analysis",
#   plot_subtitle = "Filtered for intersections with ≥8 patients"
# )

Troubleshooting

Common Issues

Empty Intersections

# Problem: Too many empty intersections
# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')
cat('This demonstrates set intersection analysis capabilities\n')

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = your_data,
#   set_vars = your_variables,
#   value_var = NULL,
#   keep_empty_groups = FALSE,  # Solution: Remove empty groups
#   min_size = 5               # Solution: Increase minimum size
# )

Too Many Sets

# Problem: Plot becomes unreadable with >8 sets
# Solution: Group related variables or create multiple plots
grouped_vars1 <- c("Surgery", "Chemotherapy", "Radiotherapy")
grouped_vars2 <- c("Immunotherapy", "TargetedTherapy", "Hormone_Therapy")

# Create separate plots for each group

Data Type Issues

# Problem: Variables not recognized as binary
# Solution: Convert to logical
your_data <- your_data %>%
  mutate(
    var1 = as.logical(var1),
    var2 = case_when(
      var2 == "Yes" ~ TRUE,
      var2 == "No" ~ FALSE,
      TRUE ~ NA
    )
  )

Advanced Examples

Stratified Analysis

# Separate analysis by patient groups
# Example: Treatment patterns by age group
young_patients <- jcomplexupset_test_data %>% filter(Age < 50)
elderly_patients <- jcomplexupset_test_data %>% filter(Age >= 65)

# Young patients
# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = young_patients,
#   set_vars = c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy"),
#   value_var = NULL,
#   min_size = 3,
#   plot_title = "Treatment Patterns in Young Patients (<50 years)",
#   plot_subtitle = paste("N =", nrow(young_patients), "patients")
# )

Time-Series Analysis

# Example of temporal analysis setup
# (Would require time-stamped data in practice)
cat("For temporal UpSet analysis, consider:\n")

## For temporal UpSet analysis, consider:

cat("- Creating time windows (quarters, years)\n")

## - Creating time windows (quarters, years)

cat("- Analyzing treatment evolution over time\n")

## - Analyzing treatment evolution over time

cat("- Comparing guideline adherence across periods\n")

## - Comparing guideline adherence across periods

Reporting Results

Statistical Summary

# Generate summary statistics for reporting
treatment_vars <- c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy")
treatment_summary <- jcomplexupset_test_data[, treatment_vars]

# Individual treatment frequencies
individual_freq <- colSums(treatment_summary)
cat("Individual treatment frequencies:\n")

## Individual treatment frequencies:

print(individual_freq)

##       Surgery  Chemotherapy  Radiotherapy Immunotherapy 
##           140           122           107            57

# Most common combinations
combinations <- treatment_summary %>%
  group_by(across(all_of(treatment_vars))) %>%
  summarise(count = n(), .groups = "drop") %>%
  arrange(desc(count)) %>%
  filter(count > 5)

cat("\nMost common treatment combinations (n > 5):\n")

## 
## Most common treatment combinations (n > 5):

print(combinations)

## # A tibble: 13 × 5
##    Surgery Chemotherapy Radiotherapy Immunotherapy count
##    <lgl>   <lgl>        <lgl>        <lgl>         <int>
##  1 TRUE    TRUE         TRUE         FALSE            32
##  2 TRUE    TRUE         FALSE        FALSE            28
##  3 TRUE    FALSE        TRUE         FALSE            21
##  4 TRUE    FALSE        FALSE        FALSE            19
##  5 TRUE    TRUE         TRUE         TRUE             15
##  6 FALSE   TRUE         FALSE        FALSE            14
##  7 FALSE   TRUE         TRUE         FALSE            13
##  8 TRUE    TRUE         FALSE        TRUE             10
##  9 FALSE   FALSE        TRUE         FALSE             9
## 10 TRUE    FALSE        FALSE        TRUE              8
## 11 FALSE   FALSE        FALSE        FALSE             7
## 12 FALSE   TRUE         TRUE         TRUE              7
## 13 TRUE    FALSE        TRUE         TRUE              7

Publication-Ready Figures

# High-quality figure for publication
# Note: This section demonstrates jcomplexupset functionality
# Due to ComplexUpset package dependency issues, the actual function call is commented out
# The jcomplexupset function would create an UpSet plot here

cat('UpSet plot would be displayed here with jcomplexupset function\n')

## UpSet plot would be displayed here with jcomplexupset function

cat('This demonstrates set intersection analysis capabilities\n')

## This demonstrates set intersection analysis capabilities

# Original function call (commented out due to dependency issues):
# jcomplexupset(
#   data = jcomplexupset_test_data,
#   set_vars = c("Surgery", "Chemotherapy", "Radiotherapy", "Immunotherapy"),
#   value_var = NULL,
#   min_size = 8,
#   sort_by = "cardinality",
#   theme_style = "theme_classic",
#   color_palette = "Set2",
#   text_size = 12,
#   show_percentages = FALSE,
#   plot_title = "",  # Remove title for publication
#   plot_subtitle = ""
# )

Conclusion

The jcomplexupset function provides a powerful and flexible tool for visualizing complex set relationships in clinical and pathological research. Key takeaways:

Use UpSet plots when analyzing 3+ sets where Venn diagrams become cluttered
Focus on clinically meaningful intersections by setting appropriate minimum sizes
Customize visualizations to match your analysis goals and audience
Combine with statistical analysis for comprehensive insights
Consider stratified analyses for different patient populations

ClinicoPath Package

2025-07-13