### Install scProportionTest R Library

Source: https://github.com/rpolicastro/scproportiontest/blob/master/README.md

Installs the scProportionTest R library from GitHub using the devtools package. Ensure devtools is installed prior to running this command.

```R
devtools::install_github("rpolicastro/scProportionTest")
```

--------------------------------

### Initialize scProportionTest Analysis Object

Source: https://github.com/rpolicastro/scproportiontest/blob/master/README.md

Loads the scProportionTest library, reads example Seurat data, and creates an analysis object using the sc_utils function. This function requires a pre-processed Seurat object as input.

```R
library("scProportionTest")

seurat_data <- system.file("extdata", "example_data.RDS", package = "scProportionTest")
seurat_data <- readRDS(seurat_data)

prop_test <- sc_utils(seurat_data)
```

--------------------------------

### Create sc_utils Object from Seurat

Source: https://context7.com/rpolicastro/scproportiontest/llms.txt

Constructs an sc_utils object by extracting cell metadata from a Seurat object. This object is essential for subsequent analyses within the scProportionTest workflow. It requires the Seurat package and a processed Seurat object as input.

```r
library("scProportionTest")
library("Seurat")

# Load example Seurat object with processed scRNA-seq data
seurat_data <- system.file("extdata", "example_data.RDS", package = "scProportionTest")
seurat_obj <- readRDS(seurat_data)

# Create sc_utils object from Seurat metadata
# Extracts all metadata columns including cluster assignments and sample identities
prop_test <- sc_utils(seurat_obj)

# The object now contains:
# - meta_data slot: data.table with cell_id and all metadata columns
# - results slot: empty list ready to store analysis results
```

--------------------------------

### Generate Point-Range Plot for Proportion Test Results

Source: https://github.com/rpolicastro/scproportiontest/blob/master/README.md

Creates a point-range plot to visualize the results of the permutation test. This function takes the analysis object containing the test results as input.

```R
permutation_plot(prop_test)
```

--------------------------------

### Visualize Proportion Test Results

Source: https://context7.com/rpolicastro/scproportiontest/llms.txt

Generates a ggplot2-based point-range plot to visualize the results of the proportion test. It displays observed log2 fold differences with bootstrap confidence intervals, color-coded by statistical significance based on FDR and log2 fold change thresholds. The plot can be customized with additional ggplot2 layers.

```r
# Create publication-ready plot of proportion test results
# FDR_threshold: significance cutoff for false discovery rate (default: 0.05)
# log2FD_threshold: effect size cutoff for biological significance (default: log2(1.5))
# order_clusters: sort clusters by observed log2 fold difference (default: TRUE)

p <- permutation_plot(
  prop_test,
  FDR_threshold = 0.05,
  log2FD_threshold = log2(1.5),
  order_clusters = TRUE
)

# The plot displays:
# - X-axis: clusters ordered by effect size
# - Y-axis: observed log2 fold difference
# - Points: observed log2FD values
# - Error bars: 95% bootstrap confidence intervals
# - Horizontal lines: zero reference (solid) and significance thresholds (dashed)
# - Colors: salmon for significant results, grey for non-significant

# Save or display the plot
ggsave("proportion_test_results.pdf", p, width = 8, height = 6)
print(p)

# Customize plot with ggplot2 layers
library(ggplot2)
p +
  labs(
    title = "Cluster Proportion Changes",
    x = "Cell Clusters",
    y = "Log2 Fold Difference (Treatment vs Control)"
  ) +
  theme(legend.position = "bottom")
```

--------------------------------

### Perform Permutation Test for Cluster Proportions

Source: https://context7.com/rpolicastro/scproportiontest/llms.txt

Conducts a permutation test to compare cell cluster proportions between two samples, also generating bootstrap confidence intervals for the log2 fold differences. It requires a sc_utils object, cluster and sample identity column names, and reference/comparison sample labels. The number of permutations influences the accuracy of p-values and confidence intervals.

```r
# Run permutation test comparing two experimental conditions
# cluster_identity: column name containing cluster assignments
# sample_1: reference sample (e.g., control)
# sample_2: comparison sample (e.g., treatment)
# sample_identity: column name containing sample labels
# n_permutations: number of permutations for both permutation test and bootstrap (default: 1000)

prop_test <- permutation_test(
  prop_test,
  cluster_identity = "custom_clusters",
  sample_1 = "HT29_EV",
  sample_2 = "HT29_LSD1_KD",
  sample_identity = "orig.ident",
  n_permutations = 1000
)

# Access results stored in prop_test@results$permutation
# Returns data.table with columns:
# - clusters: cluster names
# - HT29_EV: fraction of cells in control
# - HT29_LSD1_KD: fraction of cells in treatment
# - obs_log2FD: observed log2 fold difference
# - pval: permutation test p-value
# - FDR: false discovery rate (Benjamini-Hochberg correction)
# - boot_mean_log2FD: bootstrap mean of log2 fold difference
# - boot_CI_2.5, boot_CI_97.5: 95% confidence interval bounds

results <- prop_test@results$permutation
print(results)
# Example output:
#    clusters  HT29_EV  HT29_LSD1_KD  obs_log2FD      pval       FDR  boot_mean_log2FD  boot_CI_2.5  boot_CI_97.5
# 1: cluster1    0.150         0.220       0.552    0.023     0.069             0.548        0.251         0.842
# 2: cluster2    0.100         0.080      -0.322    0.456     0.608            -0.318       -0.721         0.089
```

--------------------------------

### Perform Permutation Test for Cell Proportions

Source: https://github.com/rpolicastro/scproportiontest/blob/master/README.md

Conducts a permutation test to compare cell proportions between specified samples within a Seurat object. Requires the analysis object, cluster identity column, sample identifiers, and the sample identity column.

```R
prop_test <- permutation_test(
	prop_test, cluster_identity = "custom_clusters",
	sample_1 = "HT29_EV", sample_2 = "HT29_LSD1_KD",
	sample_identity = "orig.ident"
)
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.