Last updated: 2022-11-10

Checks: 4 2

Knit directory: irAE_LungCancer/analysis/

This reproducible R Markdown analysis was created with workflowr (version 1.7.0). The Checks tab describes the reproducibility checks that were applied when the results were created. The Past versions tab lists the development history.

Environment: empty

Great job! The global environment was empty. Objects defined in the global environment can affect the analysis in your R Markdown file in unknown ways. For reproduciblity it’s best to always run the code in an empty environment.

Seed: set.seed(20221110)

The command set.seed(20221110) was run prior to running the code in the R Markdown file. Setting a seed ensures that any results that rely on randomness, e.g. subsampling or permutations, are reproducible.

Session information: recorded

Great job! Recording the operating system, R version, and package versions is critical for reproducibility.

Cache: detected

The following chunks had caches available:
  • unnamed-chunk-12
  • unnamed-chunk-14
  • unnamed-chunk-15
  • unnamed-chunk-18
  • unnamed-chunk-20
  • unnamed-chunk-21
  • unnamed-chunk-24
  • unnamed-chunk-26
  • unnamed-chunk-27
  • unnamed-chunk-30
  • unnamed-chunk-32
  • unnamed-chunk-33
  • unnamed-chunk-36
  • unnamed-chunk-38
  • unnamed-chunk-39

To ensure reproducibility of the results, delete the cache directory featureSelection_cache and re-run the analysis. To have workflowr automatically delete the cache directory prior to building the file, set delete_cache = TRUE when running wflow_build() or wflow_publish().

File paths: relative

Great job! Using relative paths to the files within your workflowr project makes it easier to run your code on other machines.

Repository version: no version control

Tracking code development and connecting the code version to the results is critical for reproducibility. To start using Git, open the Terminal and type git init in your project directory.

This project is not being versioned with Git. To obtain the full reproducibility benefits of using workflowr, please see ?wflow_start.

Load packages

library(MultiAssayExperiment)
library(jyluMisc)
library(tidyverse)
knitr::opts_chunk$set(warning = FALSE, message = FALSE)

Preprocessing datsets

Load data

load("../output/processedData.RData")

Only use cancer samples

mae <- mae[,mae$condition!="noMalignancy"]

CBA data

cbaMat <- mae[["cba"]]
cbaMat <- glog2(cbaMat)
mae[["cba"]] <- cbaMat

NMR data

No transformation

nmrMat <- mae[["nmr"]]

Create a new group with follow_up - baseline

CBA

cbaBase <- glog2(mae[,mae$condition == "Baseline"][["cba"]])
colnames(cbaBase) <- mae[,mae$condition == "Baseline"]$patID

cbaFollow <- glog2(mae[,mae$condition == "Follow_Up"][["cba"]])
colnames(cbaFollow) <- mae[,mae$condition == "Follow_Up"]$patID

allPat <- unique(c(colnames(cbaBase), colnames(cbaFollow)))
cbaDiff <- cbaFollow[,match(allPat,colnames(cbaFollow))] - cbaBase[,match(allPat,colnames(cbaBase))]
colnames(cbaDiff) <- allPat
cbaDiff <- cbaDiff[,complete.cases(t(cbaDiff))]

NMR

nmrBase <- mae[,mae$condition == "Baseline"][["nmr"]]
colnames(nmrBase) <- mae[,mae$condition == "Baseline"]$patID

nmrFollow <- mae[,mae$condition == "Follow_Up"][["nmr"]]
colnames(nmrFollow) <- mae[,mae$condition == "Follow_Up"]$patID

allPat <- unique(c(colnames(nmrBase), colnames(nmrFollow)))
nmrDiff <- nmrFollow[,match(allPat,colnames(nmrFollow))] - nmrBase[,match(allPat,colnames(nmrBase))]
colnames(nmrDiff) <- allPat
nmrDiff <- nmrDiff[,complete.cases(t(nmrDiff))]

Combine and create new mae object

matList <- list(cba = cbind(cbaMat, cbaDiff),
                nmr = cbind(nmrMat, nmrDiff))

#create new annotation
diffPatAnno <- colData(mae[,mae$condition != "noMalignancy"])
diffPatAnno <- diffPatAnno[!duplicated(diffPatAnno$patID),]
diffPatAnno$condition <- "Follow_Up_Baseline_diff"
rownames(diffPatAnno) <- diffPatAnno$patID
diffPatAnno <- diffPatAnno[rownames(diffPatAnno) %in% c(colnames(cbaDiff), colnames(nmrDiff)),]
newColData <- rbind(colData(mae), diffPatAnno)[,c("patID","condition","Group")]

maeNew <- MultiAssayExperiment(matList, colData = newColData)

Data completeness

plotTab <- colData(maeNew) %>% as_tibble() %>%
    mutate(present = complete.cases(t(maeNew[["cba"]]))) %>%
    filter(present)

ggplot(plotTab, aes(x=patID, y=condition, fill = present)) +
    geom_tile(col="black") +
    facet_wrap(~Group, scale = "free_x", ncol=1) +
    theme_classic()

library(randomForest)
library(caret)

Three group model

Use baseline data

Preprocessing

cbaMat <- mae[["cba"]][,mae$condition == "Baseline"]
nmrMat <- mae[["nmr"]][,mae$condition == "Baseline"]
X <- t(rbind(cbaMat, nmrMat))
y <- factor(mae[,rownames(X)]$Group)

Random forest

#Create control function for training with 10 folds and keep 3 folds for training. search method is grid.
control <- trainControl(method='repeatedcv', 
                        number=10, 
                        repeats=10, 
                        search='grid')
#create tunegrid with 15 values from 1:15 for mtry to tunning model. Our train function will change number of entry variable at each split according to tunegrid. 
tunegrid <- expand.grid(.mtry = (5:20)) 

rf_gridsearch <- train(X, y,
                       method = 'rf',
                       metric = 'Accuracy',
                       #tuneGrid = tunegrid,
                       tuneLength = 10,
                       ntree = 1000)
#print(rf_gridsearch)
plot(rf_gridsearch)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

varImpPlot(rf_gridsearch$finalModel)

SVM with non-linear kernel

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)
# Fit the model 
svms <- train(X, y, method = "svmRadial", 
              trControl = train_control, 
              preProcess = c("center","scale"), 
              tuneLength = 10)
# Print the best tuning parameter sigma and C that maximizes model accuracy
svms$bestTune
       sigma    C
1 0.01009869 0.25
plot(svms)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Multi-nomial regression (Elastic net)

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)


net.fit = train(X, y,
                method="glmnet",
                trControl=train_control,
                metric = "Accuracy",
                tuneLength = 10,
                preProcess = c("center","scale"), 
                family="multinomial")

plot(net.fit)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

impTab <- varImp(net.fit$finalModel) 
impTab <- impTab[rowSums(impTab)>0,] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature, names_to = "group", values_to = "coef") %>%
    arrange(desc(abs(coef))) %>%
    mutate(feature = factor(feature, levels = unique(feature)))
ggplot(impTab, aes(x=feature, y=coef, fill = group)) +
    geom_bar(stat = "identity", position = "dodge", width = .5) +
    theme(axis.text.x = element_text(angle = 90, hjust = 1 , vjust=0.5))

Use follw_up data

Preprocessing

cbaMat <- mae[["cba"]][,mae$condition == "Follow_Up"]
nmrMat <- mae[["nmr"]][,mae$condition == "Follow_Up"]
X <- t(rbind(cbaMat, nmrMat))
y <- factor(mae[,rownames(X)]$Group)

Random forest

#Create control function for training with 10 folds and keep 3 folds for training. search method is grid.
control <- trainControl(method='repeatedcv', 
                        number=10, 
                        repeats=10, 
                        search='grid')
#create tunegrid with 15 values from 1:15 for mtry to tunning model. Our train function will change number of entry variable at each split according to tunegrid. 
tunegrid <- expand.grid(.mtry = (5:20)) 

rf_gridsearch <- train(X, y,
                       method = 'rf',
                       metric = 'Accuracy',
                       #tuneGrid = tunegrid,
                       tuneLength = 10,
                       #preProcess = c("center","scale"), 
                       ntree = 1000)
#print(rf_gridsearch)
plot(rf_gridsearch)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

varImpPlot(rf_gridsearch$finalModel)

SVM with non-linear kernel

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)
# Fit the model 
svms <- train(X, y, method = "svmRadial", 
              trControl = train_control, 
              preProcess = c("center","scale"), 
              tuneLength = 10)
# Print the best tuning parameter sigma and C that maximizes model accuracy
svms$bestTune
       sigma C
6 0.01210822 8
plot(svms)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Multi-nomial regression (Elastic net)

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)


net.fit = train(X, y,
                method="glmnet",
                trControl=train_control,
                metric = "Accuracy",
                tuneLength = 10,
                preProcess = c("center","scale"), 
                family="multinomial")

plot(net.fit)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

impTab <- varImp(net.fit$finalModel) 
impTab <- impTab[rowSums(impTab)>0,] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature, names_to = "group", values_to = "coef") %>%
    arrange(desc(abs(coef))) %>%
    mutate(feature = factor(feature, levels = unique(feature)))
ggplot(impTab, aes(x=feature, y=coef, fill = group)) +
    geom_bar(stat = "identity", position = "dodge", width = .5) +
    theme(axis.text.x = element_text(angle = 90, hjust = 1 , vjust=0.5))

Two group model (combine rheumatic irAE and other irAE into one irAE group)

Use baseline data

Preprocessing

cbaMat <- mae[["cba"]][,mae$condition == "Baseline"]
nmrMat <- mae[["nmr"]][,mae$condition == "Baseline"]
mae$Group2 <- ifelse(mae$Group == "no irAE", mae$Group, "irAE")
X <- t(rbind(cbaMat, nmrMat))
y <- factor(mae[,rownames(X)]$Group2)

Random forest

#Create control function for training with 10 folds and keep 3 folds for training. search method is grid.
control <- trainControl(method='repeatedcv', 
                        number=10, 
                        repeats=10, 
                        search='grid')
#create tunegrid with 15 values from 1:15 for mtry to tunning model. Our train function will change number of entry variable at each split according to tunegrid. 
tunegrid <- expand.grid(.mtry = (5:20)) 

rf_gridsearch <- train(X, y,
                       method = 'rf',
                       metric = 'Accuracy',
                       tuneGrid = tunegrid,
                       ntree = 1000)
#print(rf_gridsearch)
plot(rf_gridsearch)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

varImpPlot(rf_gridsearch$finalModel)

SVM with non-linear kernel

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)
# Fit the model 
svms <- train(X, y, method = "svmRadial", 
              trControl = train_control, 
              preProcess = c("center","scale"), 
              tuneLength = 10)
# Print the best tuning parameter sigma and C that maximizes model accuracy
svms$bestTune
       sigma C
4 0.01034465 2
plot(svms)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Logistic regression (Elastic net)

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)


net.fit = train(X, y,
                method="glmnet",
                trControl=train_control,
                metric = "Accuracy",
                tuneLength = 10,
                preProcess = c("center","scale"), 
                family="binomial")

plot(net.fit)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

impTab <- varImp(net.fit$finalModel) 
impTab <- impTab[rowSums(impTab)>0,,drop=FALSE] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature, names_to = "group", values_to = "coef") %>%
    arrange(desc(abs(coef))) %>%
    mutate(feature = factor(feature, levels = unique(feature)))
ggplot(impTab, aes(x=feature, y=coef, fill = group)) +
    geom_bar(stat = "identity", position = "dodge", width = .5) +
    theme(axis.text.x = element_text(angle = 90, hjust = 1 , vjust=0.5))

Use follw_up data

Preprocessing

cbaMat <- mae[["cba"]][,mae$condition == "Follow_Up"]
nmrMat <- mae[["nmr"]][,mae$condition == "Follow_Up"]
X <- t(rbind(cbaMat, nmrMat))
y <- factor(mae[,rownames(X)]$Group2)

Random forest

#Create control function for training with 10 folds and keep 3 folds for training. search method is grid.
control <- trainControl(method='repeatedcv', 
                        number=10, 
                        repeats=10, 
                        search='grid')
#create tunegrid with 15 values from 1:15 for mtry to tunning model. Our train function will change number of entry variable at each split according to tunegrid. 
tunegrid <- expand.grid(.mtry = (5:20)) 

rf_gridsearch <- train(X, y,
                       method = 'rf',
                       metric = 'Accuracy',
                       tuneGrid = tunegrid,
                       #preProcess = c("center","scale"), 
                       ntree = 1000)
#print(rf_gridsearch)
plot(rf_gridsearch)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

varImpPlot(rf_gridsearch$finalModel)

SVM with non-linear kernel

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)
# Fit the model 
svms <- train(X, y, method = "svmRadial", 
              trControl = train_control, 
              preProcess = c("center","scale"), 
              tuneLength = 10)
# Print the best tuning parameter sigma and C that maximizes model accuracy
svms$bestTune
       sigma C
6 0.01188517 8
plot(svms)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Multi-nomial regression (Elastic net)

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)


net.fit = train(X, y,
                method="glmnet",
                trControl=train_control,
                metric = "Accuracy",
                tuneLength = 10,
                preProcess = c("center","scale"), 
                family="binomial")

plot(net.fit)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

impTab <- varImp(net.fit$finalModel) 
impTab <- impTab[rowSums(impTab)>0,,drop=FALSE] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature, names_to = "group", values_to = "coef") %>%
    arrange(desc(abs(coef))) %>%
    mutate(feature = factor(feature, levels = unique(feature)))
ggplot(impTab, aes(x=feature, y=coef, fill = group)) +
    geom_bar(stat = "identity", position = "dodge", width = .5) +
    theme(axis.text.x = element_text(angle = 90, hjust = 1 , vjust=0.5))

Combine follow_up, Baseline and Difference

Preprocessing

cbaMat <- maeNew[["cba"]] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature) %>%
    mutate(condition = colData(maeNew)[name,]$condition,
           patID = colData(maeNew)[name,]$patID) %>%
    mutate(feature = paste0(feature,"_",condition)) %>%
    select(feature, patID, value) %>%
    pivot_wider(names_from = feature, values_from = value) %>%
    column_to_rownames("patID") %>% as.matrix()
    
nmrMat <- maeNew[["nmr"]] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature) %>%
    mutate(condition = colData(maeNew)[name,]$condition,
           patID = colData(maeNew)[name,]$patID) %>%
    mutate(feature = paste0(feature,"_",condition)) %>%
    select(feature, patID, value) %>%
    pivot_wider(names_from = feature, values_from = value) %>%
    column_to_rownames("patID") %>% as.matrix()

X <- cbind(cbaMat, nmrMat)
X <- X[complete.cases(X),]

y <- factor(mae[,match(rownames(X), mae$patID)]$Group2)

Random forest

#Create control function for training with 10 folds and keep 3 folds for training. search method is grid.
control <- trainControl(method='repeatedcv', 
                        number=10, 
                        repeats=10, 
                        search='grid')
#create tunegrid with 15 values from 1:15 for mtry to tunning model. Our train function will change number of entry variable at each split according to tunegrid. 
tunegrid <- expand.grid(.mtry = (15:50)) 

rf_gridsearch <- train(X, y,
                       method = 'rf',
                       metric = 'Accuracy',
                       tuneLength = 10,
                       #tuneGrid = tunegrid,
                       #preProcess = c("center","scale"), 
                       ntree = 1000)
#print(rf_gridsearch)
plot(rf_gridsearch)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

varImpPlot(rf_gridsearch$finalModel)

SVM with non-linear kernel

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)
# Fit the model 
svms <- train(X, y, method = "svmRadial", 
              trControl = train_control, 
              preProcess = c("center","scale"), 
              tuneLength = 10)
# Print the best tuning parameter sigma and C that maximizes model accuracy
svms$bestTune
        sigma C
3 0.003072148 1
plot(svms)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Multi-nomial regression (Elastic net)

train_control <- trainControl(method="repeatedcv", number=10, repeats=10)


net.fit = train(X, y,
                method="glmnet",
                trControl=train_control,
                metric = "Accuracy",
                tuneLength = 10,
                preProcess = c("center","scale"), 
                family="binomial")

plot(net.fit)

Warning: The above code chunk cached its results, but it won’t be re-run if previous chunks it depends on are updated. If you need to use caching, it is highly recommended to also set knitr::opts_chunk$set(autodep = TRUE) at the top of the file (in a chunk that is not cached). Alternatively, you can customize the option dependson for each individual chunk that is cached. Using either autodep or dependson will remove this warning. See the knitr cache options for more details.

Feature importance

impTab <- varImp(net.fit$finalModel) 
impTab <- impTab[rowSums(impTab)>0,,drop=FALSE] %>%
    as_tibble(rownames = "feature") %>%
    pivot_longer(-feature, names_to = "group", values_to = "coef") %>%
    arrange(desc(abs(coef))) %>%
    mutate(feature = factor(feature, levels = unique(feature)))
ggplot(impTab, aes(x=feature, y=coef, fill = group)) +
    geom_bar(stat = "identity", position = "dodge", width = .5) +
    theme(axis.text.x = element_text(angle = 90, hjust = 1 , vjust=0.5))


sessionInfo()
R version 4.2.0 (2022-04-22)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Big Sur/Monterey 10.16

Matrix products: default
BLAS:   /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRblas.0.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

attached base packages:
[1] stats4    stats     graphics  grDevices utils     datasets  methods  
[8] base     

other attached packages:
 [1] caret_6.0-93                lattice_0.20-45            
 [3] randomForest_4.7-1.1        forcats_0.5.1              
 [5] stringr_1.4.0               dplyr_1.0.9                
 [7] purrr_0.3.4                 readr_2.1.2                
 [9] tidyr_1.2.0                 tibble_3.1.8               
[11] ggplot2_3.3.6               tidyverse_1.3.2            
[13] jyluMisc_0.1.5              MultiAssayExperiment_1.22.0
[15] SummarizedExperiment_1.26.1 Biobase_2.56.0             
[17] GenomicRanges_1.48.0        GenomeInfoDb_1.32.2        
[19] IRanges_2.30.0              S4Vectors_0.34.0           
[21] BiocGenerics_0.42.0         MatrixGenerics_1.8.1       
[23] matrixStats_0.62.0         

loaded via a namespace (and not attached):
  [1] utf8_1.2.2             shinydashboard_0.7.2   tidyselect_1.1.2      
  [4] htmlwidgets_1.5.4      grid_4.2.0             BiocParallel_1.30.3   
  [7] pROC_1.18.0            maxstat_0.7-25         munsell_0.5.0         
 [10] codetools_0.2-18       DT_0.23                future_1.27.0         
 [13] withr_2.5.0            colorspace_2.0-3       highr_0.9             
 [16] knitr_1.39             rstudioapi_0.13        ggsignif_0.6.3        
 [19] listenv_0.8.0          labeling_0.4.2         git2r_0.30.1          
 [22] slam_0.1-50            GenomeInfoDbData_1.2.8 KMsurv_0.1-5          
 [25] farver_2.1.1           rprojroot_2.0.3        parallelly_1.32.1     
 [28] vctrs_0.4.1            generics_0.1.3         TH.data_1.1-1         
 [31] ipred_0.9-13           xfun_0.31              sets_1.0-21           
 [34] R6_2.5.1               bitops_1.0-7           cachem_1.0.6          
 [37] fgsea_1.22.0           DelayedArray_0.22.0    assertthat_0.2.1      
 [40] promises_1.2.0.1       scales_1.2.0           multcomp_1.4-19       
 [43] nnet_7.3-17            googlesheets4_1.0.0    gtable_0.3.0          
 [46] globals_0.15.1         sandwich_3.0-2         workflowr_1.7.0       
 [49] timeDate_4021.106      rlang_1.0.4            splines_4.2.0         
 [52] rstatix_0.7.0          ModelMetrics_1.2.2.2   gargle_1.2.0          
 [55] broom_1.0.0            yaml_2.3.5             reshape2_1.4.4        
 [58] abind_1.4-5            modelr_0.1.8           backports_1.4.1       
 [61] httpuv_1.6.5           tools_4.2.0            lava_1.6.10           
 [64] relations_0.6-12       ellipsis_0.3.2         gplots_3.1.3          
 [67] jquerylib_0.1.4        proxy_0.4-27           Rcpp_1.0.9            
 [70] plyr_1.8.7             visNetwork_2.1.0       zlibbioc_1.42.0       
 [73] RCurl_1.98-1.7         ggpubr_0.4.0           rpart_4.1.16          
 [76] cowplot_1.1.1          zoo_1.8-10             haven_2.5.0           
 [79] cluster_2.1.3          exactRankTests_0.8-35  fs_1.5.2              
 [82] magrittr_2.0.3         data.table_1.14.2      reprex_2.0.1          
 [85] survminer_0.4.9        googledrive_2.0.0      mvtnorm_1.1-3         
 [88] hms_1.1.1              shinyjs_2.1.0          mime_0.12             
 [91] evaluate_0.15          xtable_1.8-4           readxl_1.4.0          
 [94] shape_1.4.6            gridExtra_2.3          compiler_4.2.0        
 [97] KernSmooth_2.23-20     crayon_1.5.1           htmltools_0.5.3       
[100] later_1.3.0            tzdb_0.3.0             lubridate_1.8.0       
[103] DBI_1.1.3              dbplyr_2.2.1           MASS_7.3-58           
[106] Matrix_1.4-1           car_3.1-0              cli_3.3.0             
[109] marray_1.74.0          parallel_4.2.0         gower_1.0.0           
[112] igraph_1.3.4           pkgconfig_2.0.3        km.ci_0.5-6           
[115] piano_2.12.0           recipes_1.0.1          xml2_1.3.3            
[118] foreach_1.5.2          bslib_0.4.0            hardhat_1.2.0         
[121] XVector_0.36.0         prodlim_2019.11.13     drc_3.0-1             
[124] rvest_1.0.2            digest_0.6.29          rmarkdown_2.14        
[127] cellranger_1.1.0       fastmatch_1.1-3        survMisc_0.5.6        
[130] kernlab_0.9-31         shiny_1.7.2            gtools_3.9.3          
[133] lifecycle_1.0.1        nlme_3.1-158           jsonlite_1.8.0        
[136] carData_3.0-5          limma_3.52.2           fansi_1.0.3           
[139] pillar_1.8.0           fastmap_1.1.0          httr_1.4.3            
[142] plotrix_3.8-2          survival_3.4-0         glue_1.6.2            
[145] iterators_1.0.14       glmnet_4.1-4           class_7.3-20          
[148] stringi_1.7.8          sass_0.4.2             caTools_1.18.2        
[151] e1071_1.7-11           future.apply_1.9.0