Package 'IOBR'

Description

Adds a custom mirror URL to the default mirrors for the current session. The mirror URL should be a base URL. If it contains 'github.com', it will be treated as a GitHub proxy and the relative path to IOBR releases will be appended. Otherwise, it will be treated as a direct repository path.

Usage

add_iobr_mirror(url, position = c("first", "last", "before_github"))

Arguments

Value

Invisibly returns the updated mirror list.

Examples

## Not run: 
add_iobr_mirror("https://my-mirror.com/https://github.com")

## End(Not run)

Description

Computes a risk score for each observation based on Cox proportional hazards regression or binary logistic regression. The function fits the specified model and returns the dataset with an added risk score column.

Usage

add_riskscore(
  input,
  family = c("cox", "binary"),
  target = NULL,
  time = NULL,
  status = NULL,
  vars,
  new_var_name = "riskscore"
)

Arguments

Value

Data frame identical to 'input' with an additional column containing risk scores (linear predictors for Cox models or predicted probabilities for logistic models).

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
input_data <- data.frame(
  time = rexp(100),
  status = rbinom(100, 1, 0.5),
  age = rnorm(100, 60, 10),
  score1 = rnorm(100),
  score2 = rnorm(100)
)
result <- add_riskscore(
  input_data,
  time = "time", status = "status",
  vars = c("age", "score1", "score2")
)
head(result$riskscore)

Description

Annotates an expression matrix with gene symbols using provided annotation data, filters out missing or invalid symbols, handles duplicate gene entries, and removes uninformative rows. The function supports multiple aggregation methods for resolving duplicate gene symbols.

Usage

anno_eset(
  eset,
  annotation,
  symbol = "symbol",
  probe = "probe_id",
  method = "mean"
)

Arguments

Details

The function performs the following operations:

1. Filters probes with missing symbols or labeled as '"NA_NA"'

2. Matches probes between expression set and annotation data

3. Merges annotation with expression data

4. Handles duplicate gene symbols using specified aggregation method

5. Removes rows with all zeros, all NAs, or missing values in the first column

Value

Annotated and cleaned gene expression matrix with gene symbols as row names.

Author(s)

Dongqiang Zeng

Examples

# Create a small example expression matrix
eset_mat <- matrix(runif(100), nrow = 10, ncol = 10)
rownames(eset_mat) <- paste0("Probe", 1:10)
colnames(eset_mat) <- paste0("Sample", 1:10)

# Create a matching annotation data frame
anno_df <- data.frame(
  probe_id = paste0("Probe", 1:10),
  symbol = c("Gene1", "Gene1", "Gene2", "Gene3", "Gene4",
             "Gene5", "Gene6", "Gene7", "Gene8", "Gene9")
)

# Annotate
result <- anno_eset(eset = eset_mat, annotation = anno_df)
head(result)

Description

Adds missing columns (filled with 'NA') to a secondary data frame so that its column set and order match a reference data frame. This is useful when combining data frames from different sources that should have the same structure but may be missing some columns.

Usage

assimilate_data(data_a, data_b)

Arguments

Value

Data frame 'data_b' with added missing columns (NA-filled) and reordered to match 'data_a'.

Examples

# Create reference data frame
pdata_a <- data.frame(
  A = 1:5, B = 2:6, C = 3:7, D = 4:8, E = 5:9
)

# Create data frame with subset of columns
pdata_b <- data.frame(A = 1:3, C = 4:6, E = 7:9)

# Harmonize pdata_b to match pdata_a structure
pdata_b_harmonized <- assimilate_data(data_a = pdata_a, data_b = pdata_b)
print(names(pdata_b_harmonized)) # Now has A, B, C, D, E

Description

Performs correlation analysis between a target variable and multiple feature variables. Computes correlation coefficients, p-values, and adjusts for multiple testing using the Benjamini-Hochberg method.

Usage

batch_cor(data, target, feature, method = c("spearman", "pearson", "kendall"))

Arguments

Value

Tibble containing the following columns for each feature:

sig_names

Feature name

p.value

Raw p-value

statistic

Correlation coefficient

p.adj

Adjusted p-value (Benjamini-Hochberg)

log10pvalue

Negative log10-transformed p-value

stars

Significance stars: **** p<0.0001, *** p<0.001, ** p<0.01, * p<0.05, + p<0.5

Author(s)

Dongqiang Zeng

Examples

# Create a small example dataset
set.seed(123)
data_df <- as.data.frame(matrix(runif(100 * 10), 100, 10))
colnames(data_df) <- paste0("Signature", 1:10)

# Perform batch correlation
results <- batch_cor(
  data = data_df,
  target = "Signature1",
  feature = colnames(data_df)[2:10]
)
head(results)

Description

Performs Kruskal-Wallis rank sum tests on multiple continuous features across different groups. Computes p-values, adjusts for multiple testing, and ranks features by significance.

Usage

batch_kruskal(data, group, feature = NULL, feature_manipulation = FALSE)

Arguments

Value

Tibble containing:

sig_names

Feature name

p.value

Raw p-value from Kruskal-Wallis test

statistic

Test statistic (chi-squared)

p.adj

Adjusted p-value (Benjamini-Hochberg)

log10pvalue

Negative log10-transformed p-value

stars

Significance stars: **** p<0.0001, *** p<0.001, ** p<0.01, * p<0.05, + p<0.5

group columns

Mean-centered values for each group

Author(s)

Dongqiang Zeng

Examples

# Create small example data
set.seed(123)
test_data <- data.frame(
  Gender = rep(c("Male", "Female"), each = 50),
  Signature1 = rnorm(100),
  Signature2 = rnorm(100)
)
# Test features by gender
res <- batch_kruskal(
  data = test_data,
  group = "Gender",
  feature = c("Signature1", "Signature2")
)
head(res)

Description

Computes partial correlation coefficients between multiple features and a target variable while controlling for an interference (confounding) variable. Adjusts p-values for multiple testing using the Benjamini-Hochberg method.

Usage

batch_pcc(
  input,
  interferenceid,
  target,
  features,
  method = c("pearson", "spearman", "kendall")
)

Arguments

Value

Tibble containing the following columns for each feature:

sig_names

Feature name

p.value

Raw p-value

statistic

Partial correlation coefficient

p.adj

Adjusted p-value (Benjamini-Hochberg method)

log10pvalue

Negative log10-transformed p-value

stars

Significance stars: **** p.adj<0.0001, *** p.adj<0.001, ** p.adj<0.01, * p.adj<0.05, + p.adj<0.5

Author(s)

Rongfang Shen

Examples

# Create small example data
set.seed(123)
test_data <- data.frame(
  TumorPurity = runif(100),
  TargetVar = rnorm(100),
  Signature1 = rnorm(100),
  Signature2 = rnorm(100)
)
# Calculate partial correlations controlling for tumor purity
res <- batch_pcc(
  input = test_data,
  interferenceid = "TumorPurity",
  target = "TargetVar",
  method = "pearson",
  features = c("Signature1", "Signature2")
)
head(res)

Description

Generates Kaplan-Meier survival plots for multiple projects or cohorts based on signature scores. Automatically determines optimal cutpoints for signature stratification and creates publication-ready survival curves.

Usage

batch_sig_surv_plot(
  input_pdata,
  signature,
  id = "ID",
  column_of_project = "ProjectID",
  project = NULL,
  time = "time",
  status = "status",
  time_type = "day",
  break_month = "auto",
  palette = "jama",
  cols = NULL,
  mini_sig = "score",
  save_path = NULL,
  show_col = TRUE,
  fig_type = "pdf"
)

Arguments

Value

Data frame containing combined survival analysis results from all projects.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
test_pdata <- data.frame(
  ID = paste0("S", 1:20),
  ProjectID = rep("P1", 20),
  OS_time = runif(20, 1, 60),
  OS_status = sample(c(0, 1), 20, replace = TRUE),
  Marker = rnorm(20)
)
result <- batch_sig_surv_plot(
  input_pdata = test_pdata, signature = "Marker",
  id = "ID", column_of_project = "ProjectID",
  time = "OS_time", status = "OS_status", time_type = "month"
)
if (!is.null(result)) head(result)

Description

Performs Cox proportional hazards regression analysis on multiple variables. Optionally determines optimal cutoffs to dichotomize continuous predictors before modeling. Returns hazard ratios, confidence intervals, and p-values for each variable.

Usage

batch_surv(
  pdata,
  variable,
  time = "time",
  status = "status",
  best_cutoff = FALSE
)

Arguments

Value

Data frame containing hazard ratios (HR), 95 and p-values for each variable, sorted by p-value.

Author(s)

Dongqiang Zeng

Examples

# Create small example data
set.seed(123)
test_data <- data.frame(
  OS_time = runif(100, 0, 100),
  OS_status = sample(c(0, 1), 100, replace = TRUE),
  Signature1 = rnorm(100),
  Signature2 = rnorm(100)
)
batch_surv(
  pdata = test_data,
  variable = c("Signature1", "Signature2"),
  time = "OS_time",
  status = "OS_status"
)

Description

Performs Wilcoxon rank-sum tests (Mann-Whitney U tests) to compare the distribution of specified features between two groups. Computes p-values, adjusts for multiple testing, and ranks features by significance.

Usage

batch_wilcoxon(
  data,
  target = "group",
  feature = NULL,
  feature_manipulation = FALSE
)

Arguments

Value

Tibble with columns:

sig_names

Feature name

p.value

Raw p-value

statistic

Difference in means between groups

p.adj

Adjusted p-value (Benjamini-Hochberg)

log10pvalue

Negative log10-transformed p-value

stars

Significance stars: **** p<0.0001, *** p<0.001, ** p<0.01, * p<0.05, + p<0.5

group1, group2

Mean values for each group

Author(s)

Dongqiang Zeng

Examples

# Create small example data
set.seed(123)
test_data <- data.frame(
  Gender = rep(c("Male", "Female"), each = 50),
  Signature1 = rnorm(100),
  Signature2 = rnorm(100)
)
# Compare features by gender
res <- batch_wilcoxon(
  data = test_data,
  target = "Gender",
  feature = c("Signature1", "Signature2")
)
head(res)

Description

Determines the optimal cutoff point for a continuous variable in survival analysis using the maximally selected rank statistics method. Creates a binary variable based on the identified cutoff and adds it to the input data frame.

Usage

best_cutoff(
  pdata,
  variable,
  time = "time",
  status = "status",
  print_result = TRUE
)

Arguments

Value

Data frame identical to 'pdata' with an additional binary column named '<variable>_binary' containing "High" and "Low" categories based on the optimal cutoff.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
pdata <- data.frame(
  time = rexp(100),
  status = rbinom(100, 1, 0.5),
  score = rnorm(100, mean = 50, sd = 10)
)
result <- best_cutoff(pdata, variable = "score", print_result = FALSE)
table(result$score_binary)

Description

Finds the best cutoff point for a continuous variable in survival analysis. Takes input data containing a continuous variable and survival information (time and status). Returns the modified input data with a new binary variable created based on the best cutoff point.

Usage

best_cutoff2(
  pdata,
  variable,
  time = "time",
  status = "status",
  print_result = TRUE
)

Arguments

Value

List containing:

pdata

Data frame with binary variable added

best_cutoff

Numeric cutoff value

cox_continuous_object

Cox model summary for continuous variable

summary_binary_variable

Summary of binary variable

cox_binary_object

Cox model summary for binary variable

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
pdata <- data.frame(
  time = rexp(100),
  status = rbinom(100, 1, 0.5),
  score = rnorm(100, mean = 50, sd = 10)
)
result <- best_cutoff2(pdata, variable = "score", print_result = FALSE)
result$best_cutoff

Description

Computes Area Under the ROC Curve (AUC) for model predictions using the ROCR package. Handles binary classification models from glmnet.

Usage

BinomialAUC(model, newx, s, acture.y)

Arguments

Value

Numeric AUC value between 0 and 1.

Examples

if (requireNamespace("glmnet", quietly = TRUE) &&
  requireNamespace("ROCR", quietly = TRUE)) {
  set.seed(123)
  train_data <- matrix(rnorm(100 * 5), ncol = 5)
  train_outcome <- rbinom(100, 1, 0.5)
  test_data <- matrix(rnorm(50 * 5), ncol = 5)
  test_outcome <- rbinom(50, 1, 0.5)
  fitted_model <- glmnet::cv.glmnet(train_data, train_outcome, family = "binomial", nfolds = 5)
  auc_value <- BinomialAUC(fitted_model, test_data, fitted_model$lambda.min, test_outcome)
  print(auc_value)
}

Description

Constructs and evaluates binomial logistic regression models using Lasso and Ridge regularization. Processes input data, scales features if specified, splits data into training/testing sets, and fits both Lasso and Ridge models. Optionally generates AUC plots for model evaluation.

Usage

BinomialModel(
  x,
  y,
  seed = 123456,
  scale = TRUE,
  train_ratio = 0.7,
  nfold = 10,
  plot = TRUE,
  palette = "jama",
  cols = NULL
)

Arguments

Value

List containing:

lasso_result

Lasso model results

ridge_result

Ridge model results

train.x

Training data with IDs

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
x <- data.frame(
  ID = paste0("Sample", 1:50),
  Feature1 = rnorm(50),
  Feature2 = rnorm(50),
  Feature3 = rnorm(50)
)
y <- data.frame(
  ID = x$ID,
  Outcome = factor(rbinom(50, 1, 0.5))
)
result <- BinomialModel(x = x, y = y, plot = FALSE, nfold = 5)
str(result, max.level = 1)

Description

Divides time duration into specified blocks for analysis.

Usage

calculate_break_month(input, block = 6, time_type = c("month", "day"))

Arguments

Value

Numeric vector of breakpoints, rounded to nearest multiple of 5.

Author(s)

Dongqiang Zeng

Examples

time_data <- c(24, 36, 12, 48)
blocks <- calculate_break_month(input = time_data)

Description

Main interface for calculating signature scores from gene expression data. Supports PCA, z-score, ssGSEA, and integration methods.

Usage

calculate_sig_score(
  pdata = NULL,
  eset,
  signature = NULL,
  method = c("pca", "ssgsea", "zscore", "integration"),
  mini_gene_count = 3,
  column_of_sample = "ID",
  print_gene_proportion = FALSE,
  print_filtered_signatures = FALSE,
  adjust_eset = FALSE,
  parallel.size = 1L,
  ...
)

Arguments

Value

Tibble with phenotype data and signature scores.

Author(s)

Dongqiang Zeng

References

1. Hänzelmann S, Castelo R, Guinney J. GSVA: gene set variation analysis. BMC Bioinformatics. 2013;14:7.

2. Mariathasan S, et al. TGF$\beta$ attenuates tumour response to PD-L1 blockade. Nature. 2018;554:544-548.

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
signature <- list(
  Signature1 = paste0("Gene", 1:10),
  Signature2 = paste0("Gene", 11:20)
)

result <- calculate_sig_score(eset = eset, signature = signature, method = "pca")

Description

Computes signature scores using PCA, z-score, and ssGSEA methods combined.

Usage

calculate_sig_score_integration(
  pdata = NULL,
  eset,
  signature,
  mini_gene_count = 2,
  column_of_sample = "ID",
  adjust_eset = FALSE,
  parallel.size = 1L
)

Arguments

Value

Tibble with signature scores from all three methods.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
signature <- list(
  Signature1 = paste0("Gene", 1:15),
  Signature2 = paste0("Gene", 16:30)
)

result <- calculate_sig_score_integration(eset = eset, signature = signature)

Description

Computes signature scores using Principal Component Analysis. The first principal component is used as the signature score.

Usage

calculate_sig_score_pca(
  pdata = NULL,
  eset,
  signature,
  mini_gene_count = 3,
  column_of_sample = "ID",
  adjust_eset = FALSE
)

Arguments

Value

Tibble with signature scores.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
signature <- list(
  Signature1 = paste0("Gene", 1:10),
  Signature2 = paste0("Gene", 11:20)
)

result <- calculate_sig_score_pca(eset = eset, signature = signature)

Description

Computes signature scores using single-sample Gene Set Enrichment Analysis.

Usage

calculate_sig_score_ssgsea(
  pdata = NULL,
  eset,
  signature,
  mini_gene_count = 5,
  column_of_sample = "ID",
  adjust_eset = FALSE,
  parallel.size = 1L
)

Arguments

Value

Tibble with signature scores.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
signature <- list(
  Signature1 = paste0("Gene", 1:15),
  Signature2 = paste0("Gene", 16:30)
)

result <- calculate_sig_score_ssgsea(eset = eset, signature = signature)

Description

Computes signature scores using z-score transformation.

Usage

calculate_sig_score_zscore(
  pdata = NULL,
  eset,
  signature,
  mini_gene_count = 3,
  column_of_sample = "ID",
  adjust_eset = FALSE
)

Arguments

Value

Tibble with signature scores.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
signature <- list(
  Signature1 = paste0("Gene", 1:10),
  Signature2 = paste0("Gene", 11:20)
)

result <- calculate_sig_score_zscore(eset = eset, signature = signature)

Description

Computes True Positive Rate (TPR) and False Positive Rate (FPR) for ROC analysis using the ROCR package. Used internally for ROC curve generation.

Usage

CalculatePref(model, newx, s, acture.y)

Arguments

Value

ROCR performance object containing TPR and FPR values.

Examples

if (requireNamespace("glmnet", quietly = TRUE) &&
  requireNamespace("ROCR", quietly = TRUE)) {
  fitted_model <- glmnet::cv.glmnet(matrix(rnorm(100), ncol = 2), rbinom(50, 1, 0.5), nfolds = 3)
  perf <- CalculatePref(fitted_model, matrix(rnorm(20), ncol = 2), "lambda.min", rbinom(10, 1, 0.5))
}

Description

Computes time-dependent ROC curve for survival models using the 'timeROC' package. Evaluates predictive accuracy at a specified time quantile.

Usage

CalculateTimeROC(model, newx, s, acture.y, modelname, time_prob = 0.9)

Arguments

Value

An object of class 'timeROC' containing ROC curve information.

Author(s)

Dongqiang Zeng

Examples

if (requireNamespace("glmnet", quietly = TRUE) &&
  requireNamespace("survival", quietly = TRUE) &&
  requireNamespace("timeROC", quietly = TRUE)) {
  library(survival)
  dat <- na.omit(lung[, c("time", "status", "age", "sex", "ph.ecog")])
  dat$status <- dat$status - 1
  x <- as.matrix(dat[, c("age", "sex", "ph.ecog")])
  y <- Surv(dat$time, dat$status)
  fit <- glmnet::glmnet(x, y, family = "cox")
  actual_outcome <- data.frame(time = dat$time, status = dat$status)
  roc_info <- CalculateTimeROC(
    model = fit, newx = x, s = 0.01, acture.y = actual_outcome,
    modelname = "glmnet Cox Model", time_prob = 0.5
  )
  print(roc_info$AUC)
}

Description

Creates stacked bar charts to visualize tumor microenvironment (TME) cell fractions. Supports batch visualization of deconvolution results from methods such as CIBERSORT, EPIC, and quanTIseq.

Usage

cell_bar_plot(
  input,
  id = "ID",
  title = "Cell Fraction",
  features = NULL,
  pattern = NULL,
  legend.position = "bottom",
  coord_flip = TRUE,
  palette = 3,
  show_col = FALSE,
  cols = NULL
)

Arguments

Value

A ggplot2 object representing the stacked bar chart.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
input_data <- data.frame(
  ID = paste0("Sample", 1:10),
  Cell_A = runif(10, 0, 0.4),
  Cell_B = runif(10, 0, 0.3),
  Cell_C = runif(10, 0, 0.3)
)
cell_bar_plot(input = input_data, id = "ID", features = c("Cell_A", "Cell_B", "Cell_C"))

Description

Processes input data containing cancer types and validates each category against a predefined list of supported cancer types ('timer_available_cancers').

Usage

check_cancer_types(args)

Arguments

Value

A character matrix with two columns:

Column 1

Expression identifiers

Column 2

Cancer categories

Examples

args <- list(
  expression = c("exp1", "exp2"),
  category = c("luad", "brca"),
  batch = NULL
)
result <- check_cancer_types(args)

Description

Performs quality checks on an expression matrix to identify missing values, infinite values, and features with zero variance. Issues warnings when potential problems are detected that may affect downstream analyses.

Usage

check_eset(eset, print_result = FALSE, estimate_sd = FALSE)

Arguments

Value

Invisibly returns 'NULL'. Function is called for its side effects (printing messages and issuing warnings).

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
check_eset(eset, print_result = TRUE)

Description

An analytical tool to estimate cell type abundances in mixed cell populations using gene expression data.

Usage

CIBERSORT(
  sig_matrix = NULL,
  mixture_file,
  perm,
  QN = TRUE,
  absolute = FALSE,
  abs_method = "sig.score",
  parallel = FALSE,
  num_cores = 2,
  seed = NULL
)

Arguments

Value

A matrix object containing the estimated cibersort-cell fractions, p-values, correlation coefficients, and RMSE values.

Author(s)

Aaron M. Newman, Stanford University ([email protected])

Examples

# Simulate data
set.seed(123)
sim_sig <- matrix(rnorm(100 * 5), 100, 5)
rownames(sim_sig) <- paste0("Gene", 1:100)
colnames(sim_sig) <- paste0("Cell", 1:5)

sim_mixture <- as.data.frame(matrix(
  rnorm(100 * 3), 100, 3
))
rownames(sim_mixture) <- paste0("Gene", 1:100)
colnames(sim_mixture) <- paste0("Sample", 1:3)

# Run deconvolution
result <- CIBERSORT(
  sig_matrix = sim_sig,
  mixture_file = sim_mixture,
  perm = 10, QN = FALSE, absolute = FALSE,
  parallel = FALSE
)
if (!is.null(result)) head(result)

Description

Removes all cached data files downloaded from GitHub.

Usage

clear_iobr_cache(cache_dir = NULL)

Arguments

Value

Invisible NULL. Called for side effects of clearing the cache.

Examples

clear_iobr_cache()

Description

Merges phenotype data with an expression matrix by matching sample IDs. Optionally filters features, applies feature manipulation, and scales expression data before combining.

Usage

combine_pd_eset(
  eset,
  pdata,
  id_pdata = "ID",
  feas = NULL,
  feature_manipulation = TRUE,
  scale = TRUE,
  choose_who_when_duplicate = c("eset", "pdata")
)

Arguments

Value

Data frame combining phenotype data and (transposed) expression data, with samples in rows and features/phenotypes in columns.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
eset <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(eset) <- paste0("Gene", 1:100)
colnames(eset) <- paste0("Sample", 1:10)
pdata <- data.frame(
  ID = colnames(eset),
  group = rep(c("A", "B"), each = 5),
  age = rnorm(10, 50, 10)
)
result <- combine_pd_eset(eset = eset, pdata = pdata, scale = FALSE)
dim(result)

Description

Creates a contrast matrix for differential analysis, where each phenotype level is contrasted against all other levels combined.

Usage

Construct_con(pheno)

Arguments

Value

Square matrix with dimensions equal to the number of levels in 'pheno'. Each row represents a contrast where the corresponding level is compared against the average of others.

Examples

pheno <- factor(c("A", "B", "C", "D"))
contrast_matrix <- Construct_con(pheno)
print(contrast_matrix)

Description

Processes a gene expression data matrix by modifying its row names. Extracts the gene identifier from row names formatted as 'GENE|ID', simplifying them to 'GENE'.

Usage

ConvertRownameToLoci(cancerGeneExpression)

Arguments

Value

Matrix with modified gene expression data with updated row names. Rows without a valid identifier are removed.

Examples

example_data <- matrix(runif(20), ncol = 5)
rownames(example_data) <- c("LOC101", "LOC102", "LOC103", "LOC104")
processed_data <- ConvertRownameToLoci(example_data)
print(processed_data)

Description

Performs nu-regression using support vector machines (SVM) to estimate cell type proportions. This is the core computational engine of CIBERSORT, using nu-SVR with linear kernel to decompose mixed gene expression signals.

Usage

CoreAlg(X, y, absolute, abs_method)

Arguments

Value

List containing:

w

Estimated cell type weights/proportions

mix_rmse

Root mean squared error of the deconvolution

mix_r

Correlation coefficient between observed and predicted mixture

Examples

# Simulate data
set.seed(123)
X <- matrix(rnorm(100), nrow = 10)
rownames(X) <- paste0("Gene", 1:10)
colnames(X) <- paste0("Cell", 1:10)
y <- rnorm(10)
names(y) <- paste0("Gene", 1:10)

# Run core algorithm
result <- CoreAlg(X, y, absolute = FALSE, abs_method = "sig.score")
if (!is.null(result)) str(result)

Description

Transforms gene expression count data into Transcripts Per Million (TPM) values, normalizing for gene length and library size. Supports multiple gene ID types and can retrieve gene length information from BioMart or use local datasets.

Usage

count2tpm(
  countMat,
  idType = "Ensembl",
  org = c("hsa", "mmus"),
  source = c("local", "biomart"),
  effLength = NULL,
  id = "id",
  gene_symbol = "symbol",
  length = "eff_length",
  check_data = FALSE
)

Arguments

Value

Data frame of TPM-normalized expression values with genes in rows and samples in columns. Gene identifiers are converted to gene symbols in the output, regardless of the input 'idType'.

Author(s)

Wubing Zhang, Dongqiang Zeng, Yiran Fang

Examples

# Simulated count data
set.seed(123)
count_matrix <- matrix(
  rpois(100, lambda = 10),
  ncol = 5,
  dimnames = list(
    paste0("ENSG00000", 101:120),
    paste0("Sample", 1:5)
  )
)
# Simulated effective length data
eff_len <- data.frame(
  id = paste0("ENSG00000", 101:120),
  symbol = paste0("Gene", 1:20),
  eff_length = runif(20, 500, 5000)
)
# Transform to TPM using local gene annotation
eset <- count2tpm(
  countMat = count_matrix, effLength = eff_len,
  id = "id", length = "eff_length", gene_symbol = "symbol"
)
head(eset)

Description

Creates one to three nested folders (if not existing) under the current working directory and returns their names and absolute paths.

Usage

creat_folder(f1, f2 = NULL, f3 = NULL, return = NULL)

Arguments

Value

List with elements:

folder_name

Relative path to the created folder

abspath

Absolute path ending with '/'

Examples

creat_folder(file.path(tempdir(), "1-result"))
creat_folder(file.path(tempdir(), "1-result"), "figures", "correlation")

Description

CIBERSORT is freely available to academic users. License and binary can be obtained from https://cibersortx.stanford.edu.

Usage

deconvo_cibersort(
  eset,
  project = NULL,
  arrays = FALSE,
  perm = 1000,
  absolute = FALSE,
  abs_method = "sig.score",
  parallel = FALSE,
  num_cores = 2,
  seed = NULL
)

Arguments

Value

Data frame with CIBERSORT cell fractions. Columns suffixed with '_CIBERSORT'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
lm22 <- load_data("lm22")
if (!is.null(lm22)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(nrow(lm22) * 2), nrow(lm22), 2)
  rownames(sim_eset) <- rownames(lm22)
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_cibersort(eset = sim_eset, project = "TCGA-STAD", perm = 10)
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Estimates immune cell fractions using EPIC algorithm.

Usage

deconvo_epic(eset, project = NULL, tumor = TRUE)

Arguments

Value

Data frame with EPIC cell fraction estimates. Columns suffixed with '_EPIC'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
TRef <- load_data("TRef")
if (!is.null(TRef)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(nrow(TRef$refProfiles) * 2), nrow(TRef$refProfiles), 2)
  rownames(sim_eset) <- rownames(TRef$refProfiles)
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_epic(eset = sim_eset, project = "Example", tumor = TRUE)
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Calculates stromal, immune, and ESTIMATE scores from gene expression.

Usage

deconvo_estimate(eset, project = NULL, platform = "affymetrix")

Arguments

Value

Data frame with ESTIMATE scores. Columns suffixed with '_estimate'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
common_genes <- load_data("common_genes")
if (!is.null(common_genes)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(nrow(common_genes) * 2), nrow(common_genes), 2)
  rownames(sim_eset) <- common_genes$GeneSymbol
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_estimate(sim_eset, project = "TCGA-STAD")
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Calculates immune phenotype scores from gene expression data.

Usage

deconvo_ips(eset, project = NULL, plot = FALSE)

Arguments

Value

Data frame with IPS scores. Columns suffixed with '_IPS'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
ips_genes <- load_data("ips_gene_set")
if (!is.null(ips_genes)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(nrow(ips_genes) * 2), nrow(ips_genes), 2)
  rownames(sim_eset) <- ips_genes$GENE
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_ips(eset = sim_eset, project = "Example")
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Estimates immune cell abundances using MCP-counter.

Usage

deconvo_mcpcounter(eset, project = NULL)

Arguments

Value

Data frame with MCP-counter scores. Columns suffixed with '_MCPcounter'.

Author(s)

Dongqiang Zeng

Examples

mcp_genes <- load_data("mcp_genes")
if (!is.null(mcp_genes)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(nrow(mcp_genes) * 3), nrow(mcp_genes), 3)
  rownames(sim_eset) <- mcp_genes$`HUGO symbols`
  colnames(sim_eset) <- paste0("Sample", 1:3)
  
  # Run deconvolution
  result <- deconvo_mcpcounter(eset = sim_eset, project = "TCGA-STAD")
  if (!is.null(result)) head(result)
}

Description

quanTIseq deconvolution for RNA-seq immune cell fractions.

Usage

deconvo_quantiseq(eset, project = NULL, tumor, arrays, scale_mrna)

Arguments

Value

Data frame with quanTIseq cell fractions. Columns suffixed with '_quantiseq'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
quantiseq_data <- load_data("quantiseq_data")
if (!is.null(quantiseq_data)) {
  set.seed(123)
  n_sig <- nrow(quantiseq_data$TIL10_signature)
  sim_eset <- matrix(rnorm(n_sig * 2), n_sig, 2)
  rownames(sim_eset) <- rownames(quantiseq_data$TIL10_signature)
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_quantiseq(eset = sim_eset, project = "Example", tumor = TRUE,
                              arrays = FALSE, scale_mrna = FALSE)
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Cell fraction estimation using SVR or lsei methods with custom reference.

Usage

deconvo_ref(
  eset,
  project = NULL,
  arrays = TRUE,
  method = c("svr", "lsei"),
  perm = 100,
  reference,
  scale_reference = TRUE,
  absolute.mode = FALSE,
  abs.method = "sig.score"
)

Arguments

Value

Data frame with cell fractions. Columns suffixed with '_CIBERSORT'.

Author(s)

Dongqiang Zeng, Rongfang Shen

Examples

# Simulate data
set.seed(123)
sim_ref <- matrix(rnorm(100 * 5), 100, 5)
rownames(sim_ref) <- paste0("Gene", 1:100)
colnames(sim_ref) <- paste0("CellType", 1:5)

sim_eset <- matrix(rnorm(100 * 3), 100, 3)
rownames(sim_eset) <- paste0("Gene", 1:100)
colnames(sim_eset) <- paste0("Sample", 1:3)

# Run deconvolution
result <- deconvo_ref(eset = sim_eset, reference = sim_ref, method = "lsei")
if (!is.null(result)) head(result)

Description

TIMER deconvolution for cancer-specific immune estimation.

Usage

deconvo_timer(eset, project = NULL, indications = NULL)

Arguments

Value

Data frame with TIMER cell fractions. Columns suffixed with '_TIMER'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
immune <- load_data("immuneCuratedData")
cancer_genes <- load_data("cancer_type_genes")
if (!is.null(immune) && !is.null(cancer_genes)) {
  set.seed(123)
  genes <- unique(c(head(rownames(immune$genes), 100), cancer_genes[["stad"]]))
  sim_eset <- matrix(rnorm(length(genes) * 2), length(genes), 2)
  rownames(sim_eset) <- genes
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_timer(eset = sim_eset, project = "TCGA-STAD",
                          indications = rep("stad", 2))
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Unified interface for multiple TME deconvolution methods.

Usage

deconvo_tme(
  eset,
  project = NULL,
  method = tme_deconvolution_methods,
  arrays = FALSE,
  tumor = TRUE,
  perm = 1000,
  reference,
  scale_reference = TRUE,
  plot = FALSE,
  scale_mrna = TRUE,
  group_list = NULL,
  platform = "affymetrix",
  absolute.mode = FALSE,
  abs.method = "sig.score",
  ...
)

Arguments

Value

Tibble with cell fractions and 'ID' column.

Author(s)

Dongqiang Zeng, Rongfang Shen

References

1. Newman et al. (2015). Robust enumeration of cell subsets from tissue expression profiles. Nature Methods.

2. Vegesna et al. (2013). Inferring tumour purity and stromal/immune cell admixture. Nature Communications.

3. Finotello et al. (2019). Molecular and pharmacological modulators of the tumor immune contexture. Genome Medicine.

4. Li et al. (2016). Comprehensive analyses of tumor immunity. Genome Biology.

5. Charoentong et al. (2017). Pan-cancer Immunogenomic Analyses. Cell Reports.

6. Becht et al. (2016). Estimating population abundance of tissue-infiltrating immune cells. Genome Biology.

7. Aran et al. (2017). xCell: digitally portraying tissue cellular heterogeneity. Genome Biology.

8. Racle et al. (2017). Simultaneous enumeration of cancer and immune cell types. ELife.

Examples

mcp_genes <- load_data("mcp_genes")
if (!is.null(mcp_genes)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(nrow(mcp_genes) * 3), nrow(mcp_genes), 3)
  rownames(sim_eset) <- mcp_genes$`HUGO symbols`
  colnames(sim_eset) <- paste0("Sample", 1:3)
  
  # Run deconvolution
  result <- deconvo_tme(eset = sim_eset, method = "mcpcounter")
  if (!is.null(result)) head(result)
}

Description

Estimates immune cell fractions using the xCell algorithm. xCell provides cell type enrichment scores for 64 immune and stromal cell types from gene expression data.

Usage

deconvo_xcell(eset, project = NULL, arrays = FALSE)

Arguments

Value

Data frame with xCell enrichment scores. Cell type columns are suffixed with '_xCell'.

Author(s)

Dongqiang Zeng

Examples

## Not run: 
xcell <- load_data("xCell.data")
if (!is.null(xcell)) {
  set.seed(123)
  sim_eset <- matrix(rnorm(length(xcell$genes) * 2), length(xcell$genes), 2)
  rownames(sim_eset) <- xcell$genes
  colnames(sim_eset) <- paste0("Sample", 1:2)
  result <- deconvo_xcell(eset = sim_eset, project = "TCGA-STAD")
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Deconvolutes gene expression data to estimate immune cell fractions using the quanTIseq method. Source code from https://github.com/FFinotello/quanTIseq.

Usage

deconvolute_quantiseq.default(
  mix.mat,
  arrays = FALSE,
  signame = "TIL10",
  tumor = FALSE,
  mRNAscale = TRUE,
  method = c("lsei", "hampel", "huber", "bisquare"),
  rmgenes = "unassigned"
)

Arguments

Value

Data frame with cell fractions for each sample.

Author(s)

Finotello F, et al. (adapted for IOBR)

References

F. Finotello, C. Mayer, C. Plattner, G. Laschober, D. Rieder, H. Hackl, A. Krogsdam, W. Posch, D. Wilflingseder, S. Sopper, M. Jsselsteijn, D. Johnsons, Y. Xu, Y. Wang, M. E. Sanders, M. V. Estrada, P. Ericsson-Gonzalez, J. Balko, N. F. de Miranda, Z. Trajanoski. "quanTIseq: quantifying immune contexture of human tumors". bioRxiv 223180. https://doi.org/10.1101/223180.

Examples

## Not run: 
quantiseq_data <- load_data("quantiseq_data")
if (!is.null(quantiseq_data)) {
  common_genes <- rownames(quantiseq_data$TIL10_signature)
  tpm_matrix <- as.data.frame(matrix(
    abs(rnorm(length(common_genes) * 2, mean = 5, sd = 2)),
    nrow = length(common_genes), ncol = 2
  ))
  rownames(tpm_matrix) <- common_genes
  colnames(tpm_matrix) <- paste0("Sample", 1:2)
  results <- deconvolute_quantiseq.default(mix.mat = tpm_matrix)
  if (!is.null(results)) head(results)
}

## End(Not run)

Description

Performs deconvolution of the tumor microenvironment using the TIMER algorithm. Processes multiple cancer datasets, removes batch effects, and estimates immune cell type abundances.

Usage

deconvolute_timer.default(args)

Arguments

Value

Matrix of abundance scores for different immune cell types across multiple cancer samples.

Examples

## Not run: 
set.seed(123)
immune <- load_data("immuneCuratedData")
cancer_genes <- load_data("cancer_type_genes")
if (!is.null(immune) && !is.null(cancer_genes)) {
  gene_names <- unique(c(head(rownames(immune$genes), 30), cancer_genes[["stad"]]))
  sample_names <- paste0("Sample", 1:2)
  n_genes <- length(gene_names)
  expr <- matrix(runif(n_genes * 2, 1, 100), n_genes, 2,
                 dimnames = list(gene_names, sample_names))
  tf <- tempfile(fileext = ".tsv")
  write.table(as.data.frame(expr), tf, sep = "\t", quote = FALSE)
  batch_tf <- tempfile(fileext = ".csv")
  write.table(data.frame(path = tf, type = "stad"), batch_tf,
              sep = ",", col.names = FALSE, row.names = FALSE, quote = FALSE)
  args <- list(outdir = tempdir(), batch = batch_tf)
  result <- deconvolute_timer.default(args)
  if (!is.null(result)) head(result)
}

## End(Not run)

Description

Creates a customized ggplot2 theme based on user-specified parameters for plot elements such as title size, axis sizes, legend settings, and theme style. Supports various base themes and allows fine-tuning of visual aspects.

Usage

design_mytheme(
  theme = c("light", "bw", "classic", "classic2"),
  plot_title_size = 2,
  axis_title_size = 2,
  axis_text_size = 12,
  axis_angle = 60,
  hjust = 1,
  legend.position = "bottom",
  legend.direction = "horizontal",
  legend.size = 0.25,
  legend.key.height = 0.5,
  legend.key.width = 0.5,
  legend.size.text = 10,
  legend.box = "horizontal"
)

Arguments

Value

A ggplot2 theme object.

Author(s)

Dongqiang Zeng

Examples

library(ggplot2)
p <- ggplot(mtcars, aes(wt, mpg)) +
  geom_point()
mytheme <- design_mytheme(theme = "bw", plot_title_size = 1.5, axis_text_size = 14)
p + mytheme + ggtitle("Example Plot")

Description

Performs permutation-based sampling to generate an empirical null distribution of correlation coefficients for p-value calculation in CIBERSORT analysis. Randomly samples from the mixture data to create null distributions.

Usage

doPerm(perm, X, Y, absolute, abs_method, seed = NULL)

Arguments

Value

List containing:

dist

Numeric vector of correlation coefficients from permutations

Examples

# Simulate data
set.seed(123)
X <- matrix(rnorm(100), nrow = 10)
rownames(X) <- paste0("Gene", 1:10)
colnames(X) <- paste0("Cell", 1:10)
Y <- rnorm(10)
names(Y) <- paste0("Gene", 1:10)

# Run permutation test
result <- doPerm(10, X, Y, absolute = FALSE, abs_method = "sig.score")
if (!is.null(result)) str(result)

Description

Downloads large datasets from GitHub releases to avoid CRAN size limits. Supports multiple download mirrors for users in different regions. Data is cached locally after first download. Cache directory can be customized via 'options(IOBR.cache_dir = "your/path")'.

Usage

download_iobr_data(
  name,
  force = FALSE,
  verbose = TRUE,
  mirrors = get_default_mirrors(),
  cache_dir = NULL
)

Arguments

Value

The requested dataset.

Examples

## Not run: 
tcga_sig <- download_iobr_data("tcga_stad_sig")

## End(Not run)

Description

Creates a quantile-quantile (QQ) plot to compare gene expression distributions between cancer and immune samples. Points along the diagonal indicate similar distributions.

Usage

DrawQQPlot(cancer.exp, immune.exp, name = "")

Arguments

Value

Generates a QQ plot.

Examples

cancer_exp <- rnorm(100, mean = 5, sd = 1.5)
immune_exp <- rnorm(100, mean = 5, sd = 1.5)
DrawQQPlot(
  cancer.exp = cancer_exp,
  immune.exp = immune_exp,
  name = "Comparison of Gene Expression"
)

Description

Fits elastic net model with cross-validation to find optimal alpha and lambda. Searches across a grid of alpha values (0 to 1) and lambda values to minimize cross-validation error.

Usage

Enet(train.x, train.y, lambdamax, nfold = 10)

Arguments

Value

List containing:

chose_alpha

Optimal alpha value (0-1)

chose_lambda

Optimal lambda value

Examples

if (requireNamespace("glmnet", quietly = TRUE)) {
  set.seed(123)
  train_data <- matrix(rnorm(50 * 5), ncol = 5)
  train_outcome <- rbinom(50, 1, 0.5)
  result <- Enet(train.x = train_data, train.y = train_outcome, lambdamax = 1, nfold = 5)
}

Description

Creates a bar plot visualizing enrichment results for up-regulated and down-regulated terms, using -log10(p-values) to indicate significance.

Usage

enrichment_barplot(
  up_terms,
  down_terms,
  terms = "Description",
  pvalue = "pvalue",
  group = "group",
  palette = "jama",
  cols = NULL,
  title = "Gene Ontology Enrichment",
  width_wrap = 30,
  font_terms = 15
)

Arguments

Value

A ggplot object of the enrichment bar plot.

Author(s)

Dongqiang Zeng

Examples

up_terms <- data.frame(
  Description = c("Pathway1", "Pathway2"),
  pvalue = c(0.001, 0.01)
)
down_terms <- data.frame(
  Description = c("Pathway4", "Pathway5"),
  pvalue = c(0.005, 0.02)
)
p <- enrichment_barplot(
  up_terms = up_terms,
  down_terms = down_terms,
  title = "Custom Enrichment Plot"
)
p

Description

EPIC takes as input bulk gene expression data (RNA-seq) and returns the proportion of mRNA and cells composing the various samples.

Usage

EPIC(
  bulk,
  reference = NULL,
  mRNA_cell = NULL,
  mRNA_cell_sub = NULL,
  sigGenes = NULL,
  scaleExprs = TRUE,
  withOtherCells = TRUE,
  constrainedSum = TRUE,
  rangeBasedOptim = FALSE
)

Arguments

Details

This function uses a constrained least square minimization to estimate the proportion of each cell type with a reference profile and another uncharacterized cell type in bulk gene expression samples.

The names of the genes in the bulk samples, the reference samples and in the gene signature list need to be the same format (gene symbols are used in the predefined reference profiles). The full list of gene names don't need to be exactly the same between the reference and bulk samples: EPIC will use the intersection of the genes. In case of duplicate gene names, EPIC will use the median value per duplicate - if you want to consider these cases differently, you can remove the duplicates before calling EPIC.

Value

A list of 3 matrices: - 'mRNAProportions': (nSamples x (nCellTypes+1)) the proportion of mRNA coming from all cell types with a ref profile + the uncharacterized other cell - 'cellFractions': (nSamples x (nCellTypes+1)) this gives the proportion of cells from each cell type after accounting for the mRNA / cell value - 'fit.gof': (nSamples x 12) a matrix telling the quality for the fit of the signature genes in each sample. It tells if the minimization converged, and other info about this fit comparing the measured gene expression in the sigGenes vs predicted gene expression in the sigGenes

Examples

melanoma_counts <- matrix(abs(rnorm(80)), nrow = 20, ncol = 4)
rownames(melanoma_counts) <- paste0("Gene", 1:20)
colnames(melanoma_counts) <- paste0("Sample", 1:4)
mock_ref <- list(
  refProfiles = matrix(abs(rnorm(80)), nrow = 20, ncol = 4),
  sigGenes = paste0("Gene", 1:10)
)
rownames(mock_ref$refProfiles) <- paste0("Gene", 1:20)
colnames(mock_ref$refProfiles) <- c("Bcells", "CD4T", "CD8T", "NK")
res1 <- EPIC(melanoma_counts, reference = mock_ref)
if (!is.null(res1)) head(res1$cellFractions)

Description

Generates boxplots and density plots to analyze the distribution of expression values in an expression set. Useful for quality control and assessing data normalization.

Usage

eset_distribution(eset, quantile = 3, log = TRUE, project = NULL)

Arguments

Value

Invisibly returns 'NULL'. If 'project' is provided, saves PNG files to disk.

Examples

# Simulate data
set.seed(123)
sim_eset <- matrix(rnorm(1000 * 10, mean = 5, sd = 2), 1000, 10)
rownames(sim_eset) <- paste0("Gene", 1:1000)
colnames(sim_eset) <- paste0("Sample", 1:10)

# Run distribution plot
result <- eset_distribution(sim_eset)

Description

This function reads a gene expression dataset in GCT format, calculates enrichment scores for specific gene sets, and writes the computed scores to an output file. It supports multiple platform types and performs platform-specific calculations if necessary.

Usage

estimateScore(
  input.ds,
  output.ds,
  platform = c("affymetrix", "agilent", "illumina")
)

Arguments

Value

This function does not return a value but writes the computed scores to the specified output file in GCT format.

Examples

## Not run: 
set.seed(123)
si_geneset_data <- load_data("SI_geneset")
if (!is.null(si_geneset_data)) {
  gene_names <- unique(c(si_geneset_data[1, -1], si_geneset_data[2, -1]))
  gene_names <- gene_names[!is.na(gene_names) & gene_names != ""]
  gene_names <- head(gene_names, 200)
  n_genes <- length(gene_names)
  eset_sim <- as.data.frame(matrix(rnorm(n_genes * 2, mean = 5, sd = 1), n_genes, 2))
  rownames(eset_sim) <- gene_names
  colnames(eset_sim) <- c("Sample1", "Sample2")
  eset_sim <- tibble::rownames_to_column(eset_sim, var = "symbol")
  input_file <- tempfile(pattern = "estimate_", fileext = ".gct")
  output_file <- tempfile(pattern = "estimate_score_", fileext = ".gct")
  writeLines(c("#1.2", paste(nrow(eset_sim), ncol(eset_sim) - 1, sep = "\t")), input_file)
  utils::write.table(eset_sim, input_file, sep = "\t", row.names = FALSE,
                    col.names = TRUE, append = TRUE, quote = FALSE)
  score_res <- estimateScore(input.ds = input_file, output.ds = output_file,
                             platform = "affymetrix")
  if (!isFALSE(score_res) && file.exists(output_file)) {
    head(read.table(output_file, skip = 2, header = TRUE, sep = "\t"))
  }
}

## End(Not run)

Description

Computes the exact p-value for the correlation between two numeric variables using a specified correlation method.

Usage

exact_pvalue(x, y, method)

Arguments

Value

Numeric value representing the exact p-value.

Author(s)

Dongqiang Zeng

Examples

# Simulate data
set.seed(123)
x <- rnorm(100)
y <- rnorm(100)
p_val <- exact_pvalue(x = x, y = y, method = "spearman")
print(p_val)

Description

Extracts and combines a data frame with cells as rows and features as columns from Seurat assay data. Supports multiple assays and optional metadata integration.

Usage

extract_sc_data(
  sce,
  vars = NULL,
  assay,
  slot = "scale.data",
  combine_meta_data = TRUE
)

Arguments

Value

Data frame with cells as rows and features as columns.

Author(s)

Dongqiang Zeng

Examples

if (requireNamespace("Seurat", quietly = TRUE)) {
  pbmc <- SeuratObject::pbmc_small
  vars <- c("PPBP", "IGLL5", "VDAC3", "CD1C", "AKR1C3")
  eset <- extract_sc_data(sce = pbmc, vars = vars, assay = "RNA")
}

Description

Filters features (variables) in a matrix or data frame by removing those with missing values, non-numeric types, infinite values, or zero variance. This is useful for preparing data for downstream statistical analyses.

Usage

feature_manipulation(
  data,
  feature = NULL,
  is_matrix = FALSE,
  print_result = FALSE
)

Arguments

Value

Character vector of feature names that pass all quality checks.

Author(s)

Dongqiang Zeng

Examples

set.seed(123)
test_data <- data.frame(
  A = c(1, 2, 3),
  B = c(1, 1, 1), # zero variance
  C = c(1, NA, 3), # missing value
  D = c("a", "b", "c") # non-numeric
)
feas <- feature_manipulation(data = test_data, feature = colnames(test_data), print_result = TRUE)
print(feas)

Description

Selects informative features using either correlation with a quantitative response or differential expression (limma) for binary/continuous responses.

Usage

feature_select(
  x,
  y,
  method = c("cor", "dif"),
  family = c("spearman", "pearson"),
  cutoff = NULL,
  padjcut = NULL
)

Arguments

Value

Character vector of selected feature names.

Examples

# Simulate data
set.seed(123)
sim_eset <- matrix(rnorm(100 * 50), 100, 50)
rownames(sim_eset) <- c("PDCD1", paste0("Gene", 2:100))
colnames(sim_eset) <- paste0("Sample", 1:50)
pd1 <- as.numeric(sim_eset["PDCD1", ])
group <- ifelse(pd1 > mean(pd1), "high", "low")

# Correlation method
pd1_cor <- feature_select(
  x = sim_eset, y = pd1, method = "cor",
  family = "pearson", padjcut = 0.05, cutoff = 0.5
)

# Differential expression method
pd1_dif <- feature_select(
  x = sim_eset, y = pd1, method = "dif",
  padjcut = 0.05, cutoff = 2
)
pd1_dif_2 <- feature_select(
  x = sim_eset, y = group,
  method = "dif", padjcut = 0.05, cutoff = 2
)

Description

This function filters and merges a dataset with a set of common genes.

Usage

filterCommonGenes(input.f, output.f, id = c("GeneSymbol", "EntrezID"))

Arguments

Value

No return value. The function writes the merged dataset to the specified output file.

Examples

## Not run: 
input_data <- data.frame(
  GeneSymbol = c("BRCA1", "TP53", "EGFR", "NOTCH1"),
  Value = c(10, 15, 8, 12),
  stringsAsFactors = FALSE
)
input_file <- tempfile(fileext = ".txt")
output_file <- tempfile(fileext = ".txt")
write.table(input_data, file = input_file, sep = "\t", row.names = TRUE, quote = FALSE)
filterCommonGenes(input_file, output_file, id = "GeneSymbol")

## End(Not run)

Description

Identifies informative marker features across groups from bulk gene expression or signature score matrices using Seurat workflows. Performs feature selection, scaling, PCA, clustering, and marker discovery.

Usage

find_markers_in_bulk(
  pdata,
  eset,
  group,
  id_pdata = "ID",
  nfeatures = 2000,
  top_n = 20,
  thresh.use = 0.25,
  only.pos = TRUE,
  min.pct = 0.25,
  npcs = 30
)

Arguments

Value

List with components: 'sce' (Seurat object), 'markers' (all markers), 'top_markers' (top markers per group).

Examples

if (requireNamespace("Seurat", quietly = TRUE) && requireNamespace("Matrix", quietly = TRUE)) {
  # Simulate data
  set.seed(123)
  sim_eset <- matrix(abs(rnorm(100 * 30)), 100, 30)
  rownames(sim_eset) <- paste0("Gene", 1:100)
  colnames(sim_eset) <- paste0("Sample", 1:30)
  
  sim_pdata <- data.frame(
    ID = paste0("Sample", 1:30),
    TMEcluster = rep(c("A", "B", "C"), each = 10)
  )
  
  res <- find_markers_in_bulk(
    pdata = sim_pdata, eset = sim_eset,
    group = "TMEcluster", npcs = 5
  )
  if (!is.null(res)) head(res$top_markers)
}

Description

This function identifies mutations associated with a specific signature score, performs statistical tests for significance, and generates oncoprints and box plots to visualize relationships.

Usage

find_mutations(
  mutation_matrix,
  signature_matrix,
  id_signature_matrix = "ID",
  signature,
  min_mut_freq = 0.05,
  plot = TRUE,
  method = "multi",
  point_alpha = 0.1,
  save_path = NULL,
  palette = "jco",
  cols = NULL,
  show_plot = TRUE,
  show_col = FALSE,
  width = 8,
  height = 4,
  oncoprint_group_by = "mean",
  oncoprint_col = "#224444",
  gene_counts = 10,
  jitter = FALSE,
  genes = NULL,
  point_size = 4.5
)