data_description.md

See this paper: https://doi.org/10.1158/1055-9965.EPI-17-0461

Background: The tumor microenvironment is an important factor in cancer immunotherapy response. To further understand how a tumor affects the local immune system, we analyzed immune gene expression differences between matching normal and tumor tissue.Methods: We analyzed public and new gene expression data from solid cancers and isolated immune cell populations. We also determined the correlation between CD8, FoxP3 IHC, and our gene signatures.Results: We observed that regulatory T cells (Tregs) were one of the main drivers of immune gene expression differences between normal and tumor tissue. A tumor-specific CD8 signature was slightly lower in tumor tissue compared with normal of most (12 of 16) cancers, whereas a Treg signature was higher in tumor tissue of all cancers except liver. Clustering by Treg signature found two groups in colorectal cancer datasets. The high Treg cluster had more samples that were consensus molecular subtype 1/4, right-sided, and microsatellite-instable, compared with the low Treg cluster. Finally, we found that the correlation between signature and IHC was low in our small dataset, but samples in the high Treg cluster had significantly more CD8+ and FoxP3+ cells compared with the low Treg cluster.Conclusions: Treg gene expression is highly indicative of the overall tumor immune environment.Impact: In comparison with the consensus molecular subtype and microsatellite status, the Treg signature identifies more colorectal tumors with high immune activation that may benefit from cancer immunotherapy.

IDPT_ml.R

# Install Bioconductor and required packages
# This command installs the BiocManager package for managing Bioconductor packages
install.packages("BiocManager")

# Use BiocManager to install specific packages for data analysis and visualization
BiocManager::install(c("GEOquery", "SummarizedExperiment", "ggplot2",
                       "party", "ggparty", "partykit", "randomForest"))

# Load necessary packages for modeling and visualization
library(party)          # For creating classification trees
library(ggparty)        # For visualizing classification trees with ggplot2
library(partykit)       # Provides tools for working with party objects

# Load the iris dataset, which is a well-known dataset in R for classification
data(iris)

# Display the first few rows of the iris dataset
head(iris)

# Create a classification tree model using conditional inference trees
# The model predicts the Species based on all other variables in the dataset
py <- party::ctree(Species ~ ., data = iris)

# Visualize the classification tree created by the model
plot(py)

# Define a function to create and visualize a classification tree
plot_random_sample <- function(fraction) {
  # Create a random sample of the iris dataset
  sample_indices <- sample(1:nrow(iris), fraction * nrow(iris))  # Randomly select indices
  iris_sample <- iris[sample_indices, ]  # Create a sample dataset

  # Create a classification tree model from the random sample
  py_sample <- party::ctree(Species ~ ., data = iris_sample)

  # Visualize the classification tree for the sampled data
  plot(py_sample)
}

# Call the function to plot a classification tree using 50% of the iris dataset
plot_random_sample(0.5)
plot_random_sample(0.5)
plot_random_sample(0.5)

# Load the randomForest package for random forest modeling
library(randomForest)

# Create a random forest model to classify Species based on the iris dataset
# 'ntree = 1000' specifies that 1000 trees will be used in the model
rf = randomForest(Species ~ ., data = iris, ntree = 1000)

# Display the importance of each feature in the random forest model
importance(rf)

## Perform PCA (Principal Component Analysis)
library(GEOquery)  # Load GEOquery package for accessing GEO datasets
gse = getGEO("GSE103512")[[1]]  # Retrieve a specific GEO dataset

library(SummarizedExperiment)  # Load the package for working with summarized experiment data
se = as(gse, "SummarizedExperiment")  # Convert the GEO dataset to a SummarizedExperiment object

# Create a contingency table to show the relationship between cancer types and normal samples
with(colData(se),table(`cancer.type.ch1`,`normal.ch1`))

# Calculate the standard deviation for each gene across all samples
sds = apply(assay(se, 'exprs'), 1, sd)
# Select the top 500 genes with the highest standard deviation
dat = assay(se, 'exprs')[order(sds, decreasing = TRUE)[1:500],]

# Perform PCA on the transposed data
# 't(dat)' transposes the dataset so that samples are rows and genes are columns
# 'center = TRUE' centers the data by subtracting the mean
# 'scale. = TRUE' scales the data to have unit variance
pca_results = prcomp(t(dat), center = TRUE, scale. = TRUE)

# Convert PCA results into a data frame
# 'pca_results$x' contains the principal component scores (new coordinates for our data)
pca_vals = as.data.frame(pca_results$x)

# Add additional information to the PCA results for further analysis
# 'Type' is a factor representing different cancer types
pca_vals$Type = factor(colData(se)[,'cancer.type.ch1'])
# 'Normal' is a factor representing whether the sample is normal or not
pca_vals$Normal = factor(colData(se)[,'normal.ch1'])

# Display the first few rows of the PCA results
head(pca_vals)

# Plot the proportion of variance explained by each principal component
# This can help in determining how many components to retain for further analysis
screeplot(pca_results)

# Create a scatter plot using ggplot2 to visualize the first two principal components
# 'aes' maps the x and y axes to the first two principal components (PC1 and PC2)
# 'shape' distinguishes the points based on the Normal variable
# 'color' distinguishes the points based on the Type variable
library(ggplot2)  # Load ggplot2 for creating graphics
ggplot(pca_vals, aes(x=PC1, y=PC2, shape=Normal, color=Type)) +
  geom_point(alpha=0.6) +  # Add points to the plot with some transparency
  theme(text=element_text(size = 18))  # Increase text size for better readability