R/server.R
In EdwardAgboraw/FrankenSeq: A user friendly and modular machine learning platform for scRNA-seq analysis
Defines functions
server
Documented in
server
#'The Server Function
#'
#' @param input standard shiny parameter
#'
#' @param output standard shiny parameter
#'
#'@import magrittr
#'@import dplyr
#'@import shiny
#'@return None
server = function(input,output) {
options(shiny.maxRequestSize=300*1024^2) #increase the max file limit
#QUALITY CONTROL
rawData <- reactiveVal()
filteredData <- reactiveVal()
output$dgraph = renderPlot({
if(input$Help_QC == "Help") {
return(NULL)
}
#input raw data
rdata <- input$rdata
filePath <- input$rdata$datapath
fileType <- input$FileType
if (fileType == 'PBMC3k Test Data') {
inputData = LoadTestData()
} else {
inputData <- LoadData(fileType, filePath)
}
rawData(inputData)
#Data Processing
minNum = input$Min_Features
maxNum = input$Max_Features
percentMT = input$Mitochondria
nFeatures = input$nFeatures
cleanData <- FilterData(rawData(), minNum, maxNum, percentMT, nFeatures)
filteredData(cleanData)
if(input$DPOptions == "Quality Control") {
qcPlots <- VlnPlot(rawData(), features = c("nFeature_RNA", "nCount_RNA", "percent.mt"), ncol = 3)
return(qcPlots)
} else if (input$DPOptions == "Data Filtration") {
newQCplots <- VlnPlot(filteredData(), features = c("nFeature_RNA", "nCount_RNA", "percent.mt"), ncol = 3)
return(newQCplots)
}
})
#FEATURE SELECTION
featureGenes <- reactiveVal()# DATASET 2
featureList <- reactiveVal()
output$FSPlot = renderPlot({
if(is.null(filteredData()) || input$Help_FS == "Help") {
return(NULL)
}
fs <- input$FS
featureNumber <- input$nFeatures
fs_results <- featureSelection(filteredData(), fs, featureNumber)
featureGenes(fs_results$data)
featureList(fs_results$featurelist)
fs_plot <- fs_results$plot
return(fs_plot)
})
#DIMENSION REDUCTION
dimRData <- reactiveVal() # DATASET 3
dimRMethod <- reactiveVal()
output$drPlot = renderPlot({
if(is.null(featureGenes()) || input$Help_DR == "Help") {
return(NULL)
}
dimR <- input$dr
graphType_DR <- input$DRG_options
L <- input$L_value
if(dimR == "PCA (Seurat)") {
dr_data <- PCA_DimR(featureGenes(), graphType_DR)
dimRData(dr_data$data)
dimRMethod(dr_data$method)
return(dr_data$plot)
} else if(dimR == "GLM PCA") {
dr_data <- GLM_PCA_DimR(featureGenes(), L)
dimRData(dr_data$data)
dimRMethod(dr_data$method)
return(dr_data$plot)
} else if(dimR == "GLM PCA (Residuals)") {
dr_data <- GLM_PCA_Residuals_DimR(featureGenes(), graphType_DR)
dimRData(dr_data$data)
dimRMethod(dr_data$method)
return(dr_data$plot)
}
})
#Cluster Validation
output$cvPlot = renderPlot({
if(input$CVOptions == "Estimate K with SC3" || is.null(dimRData()) || input$Help_CV == "Help") {
return(NULL)
}
graphType_CV <- input$CVOptions
maxK <- input$maxK
cvData <- dimRData()
clusterValPlot <- clusterValidation(cvData, graphType_CV, maxK)
return(clusterValPlot)
})
#SC3 Estimate K
output$sc3_estimated_k_value <- renderText({
if(input$CVOptions != "Estimate K with SC3" || is.null(dimRData()) || input$Help_CV == "Help") {
return(NULL)
}
clusValData <- dimRData()
k <- sc3_estimate(clusValData)
string <- paste("Estimated number of clusters = ", k)
return(string)
})
#CLUSTER ANALYSIS
clusterData <- reactiveVal() # DATASET 4
output$SClusters = renderPlot({
if(is.null(dimRData()) || input$Help_CA == "Help") {
return(NULL)
}
sO <- dimRData() #this object should have both the current feature gene and reduced dimension sets already loaded.
reductionMethod <- dimRMethod() #formerly method1
clusteringMethod <- input$CM #formerly method
graphType_CA <- input$cv #formerly reductionMethod
noDim <- input$dimensions
res <- input$resolution
kValue <- input$k_value
coreNumber <- input$ncore
linkM <- input$LM
hcOption <- input$HCoptions
ha_fs_option <- input$DLoptions
ha_cl_option <- input$DLoptions2
if (clusteringMethod == "K-Nearest Neighbor (Seurat)") {
knn <- seuratClustering(sO, noDim, res, graphType_CA, reductionMethod)
clusterData(knn$data)
return(knn$plot)
} else if (clusteringMethod == "Graph Based Hierarchical Clustering (HGC)") {
hgc <- HGC_clustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(hgc$data)
return(hgc$plot)
} else if (clusteringMethod == "Hierarchical Clustering") {
hc <- h_clustering(sO, linkM, noDim, graphType_CA, reductionMethod, kValue, hcOption)
clusterData(hc$data)
return(hc$plot)
} else if (clusteringMethod == "K-means Clustering") {
km <- km_clustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(km$data)
return(km$plot)
} else if (clusteringMethod == "Density Peak Clustering (Monocle)") {
cd <- monocleClustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(cd$data)
return(cd$plot)
} else if (clusteringMethod == "Consensus Clustering (SC3)") {
cc <- consensus_clustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(cc$data)
return(cc$plot)
}
})
#OUTPUT DATA
resultsTable = reactiveVal()
output$table = renderTable({
if(is.null(clusterData()) || input$Help_R == "Help") {
return(NULL)
}
tableChoice = input$TableOptions
sO = clusterData()
resultsTable = generateTable(sO, tableChoice)
return(resultsTable)
})
output$Pipeline <- renderUI({
str1 <- paste("<B>Feature Selection Algorithm:<B> ", input$FS)
str2 <- paste("<B>Dimension Reduction Method:<B> ", input$dr)
str3 <- paste("<B>Unsupervised Clustering Method:<B> ", input$CM)
HTML(paste(str1, str2, str3, sep = '<br/>'))
})
output$dlb <- downloadHandler(
filename <- "FinalData.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(resultsTable(), file, row.names = FALSE)
}
)
output$FSdlb <- downloadHandler(
filename <- "Selected_Features.txt",
contentType <- "csv",
content <- function(file) {
utils::write.csv(featureList(), file, row.names = FALSE)
}
)
output$SOdlb <- downloadHandler(
filename <- "FinalSeuratObject.txt",
contentType<- "RDS",
content <- function(file) {
saveRDS(clusterData(), file)
}
)
#Differentially Expressed Gene Analysis
output$degPlot = renderPlot({
if(is.null(clusterData()) || input$Help_DA == "Help") {
return(NULL)
}
sO <- clusterData()
lfc <- input$LFC
minPCT <- input$minPC
outputType <- input$DEG_Options
degResults = degAnalysis(sO, lfc, minPCT, outputType)
return(degResults)
})
degMarkerTable = reactiveVal()
output$degTable = renderTable({
if(is.null(clusterData()) || input$Help_DA == "Help") {
return(NULL)
}
sO <- clusterData()
lfc <- input$LFC
minPCT <- input$minPC
outputType <- input$DEG_Options
degResults = degAnalysis(sO, lfc, minPCT, outputType)
degMarkerTable(degResults)
return(degResults)
})
output$deg_biomarkers <- downloadHandler(
filename <- "ClusterBiomarkers.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(degMarkerTable, file, row.names = FALSE)
}
)
#Deep Learning
deepLearningData <- reactiveVal()
output$DL_Plot <- renderPlot({
if(is.null(dimRData()) || input$Help_DL == "Help") {
return(NULL)
}
#Variables
sO <- dimRData()
kValue <- input$DL_Knumber
noDim <- input$DL_dimensions
coreNumber <- input$DL_ncore
reductionMethod <- dimRMethod() #formerly method1
graphType_DL <- input$DL_cv #formerly reductionMethod
#DL Function Call
deepLearningResults <- autoEncoderClusterring(sO, noDim, kValue, coreNumber, graphType_DL, reductionMethod)
deepLearningData(deepLearningResults$data)
return(deepLearningResults$plot)
})
output$dl_seurat <- downloadHandler(
filename <- "DL_SeuratObject.txt",
contentType <- "RDS",
content <- function(file) {
saveRDS(deepLearningData(), file)
}
)
resultsTable_DL = reactiveVal()
output$DL_Table = renderTable({
if(is.null(deepLearningData()) || input$Help_DL == "Help") {
return(NULL)
}
sO = deepLearningData()
tableChoice = input$DL_TableOptions
if (tableChoice == "Summary Report") {
Dsummary <- table(Idents(sO))
Dsummary <- as.data.frame(Dsummary)
colnames(Dsummary) <- c("Cluster", "Frequency")
resultsTable_DL(Dsummary)
return(Dsummary)
} else {
results <- as.data.frame(sO@active.ident)
colnames(results) <- "Cluster"
results <- cbind(rownames(results), data.frame(results, row.names = NULL))
colnames(results) <- c("Cell Label", "Cluster")
results <- results[order(results$Cluster), ]
resultsTable_DL(results)
return(results)
}
})
output$dl_table_download <- downloadHandler(
filename <- "DL_Data.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(resultsTable_DL(), file, row.names = FALSE)
}
)
output$DL_Heatmap = renderPlot({
if(is.null(deepLearningData()) || input$Help_DL == "Help") {
return(NULL)
}
#Variables
sO <- deepLearningData()
lfc_DL <- input$DL_LFC
minPCT_DL <- input$DL_minPC
degHeatMap_DL = degAnalysis(sO, lfc_DL, minPCT_DL, "Heatmap")
#sO.markers_DL <- FindAllMarkers(sO, only.pos = TRUE, min.pct = minPCT_DL, logfc.threshold = lfc_DL)
#sO.markers_DL %>% group_by(cluster) %>% top_n(n = 2, wt = avg_log2FC) -> top10
#degHeatMap_DL <- DoHeatmap(sO, features = top10$gene) + NoLegend()
return(degHeatMap_DL)
})
biomarkerTable_DL = reactiveVal()
output$DL_Biomarkers = renderTable({
if(is.null(deepLearningData()) || input$Help_DL == "Help") {
return(NULL)
}
#Variables
sO <- deepLearningData()
lfc_DL <- input$DL_LFC
minPCT_DL <- input$DL_minPC
#sO.markers_DL <- FindAllMarkers(sO, only.pos = TRUE, min.pct = dl_minpct, logfc.threshold = dl_lfc)
#sO.markers_DL %>% group_by(cluster) %>% top_n(n = 2, wt = avg_log2FC) -> top10
#degTable_DL <- dl_sO.markers %>% group_by(cluster) %>% slice_max(n = 2, order_by = avg_log2FC)
degTable_DL = degAnalysis(sO, lfc_DL, minPCT_DL, "Table")
biomarkerTable_DL(degTable_DL)
return(degTable_DL)
})
output$dl_bmTable <- downloadHandler(
filename <- "DL_BiomarkerTable.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(biomarkerTable_DL(), file, row.names = FALSE)
}
)
#The Help Text
output$QC_Help_Text <- renderText(
"This is the Data Processing Tab. Accepted inputs include Seurat/Single Cell Experiment Objects (as RDS files)
and Gene x Cell matrices (as CSV files). 'Quality Control' generates violin plots visualizing the QC metrics of the raw input data.
'Data Filtration' trims cells from the data according to user defined Feature Count and % Mitochondrial DNA limits.
Sample Test Data (PBMC3k) is also available here."
)
output$FS_Help_Text <- renderText(
"This is the Feature Selection Tab. Feature Selection identifies genes that show biologically revelent variation.
Here genes can be assessed and selected according to a variety of criteria, ranging from Variance to Drop-Out Rate."
)
output$DR_Help_Text <- renderText(
"This is the Dimension Reduction Tab. The purpose of dimension reduction is to compress input data into a 'lower dimensional' representation more compatible
with clustering. This can be achieved here using either PCA, GLM-PCA (a modification of PCA optimized for UMI count data), or Residual PCA
(a fast approximation of the GLM-PCA algorithm)."
)
output$CV_Help_Text <- renderText(
"This is the Cluster Validation Tab. This tab estimates the number of clusters in the data using either the Elbow or Silhouette Methods.
Cluster number can also be directly estimated using the Tracy-Widom Theory of Random Matrices (via SC3). This tab is both time-intensive and optional -
this part of the pipeline can safely be skipped."
)
output$CA_Help_Text <- renderText(
"This is the Cluster Analysis Tab."
)
output$R_Help_Text <- renderText(
"This is the R Help Text."
)
output$DA_Help_Text <- renderText(
"This is the Differentially Expressed Gene (DEG) Analysis Tab."
)
output$DL_Help_Text <- renderText(
"This is the Deep Learning Tab. This Tab represents a fully independent analysis pipeline."
)
}
EdwardAgboraw/FrankenSeq documentation built on May 3, 2024, 8:26 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions server
Documented in server
#'The Server Function
#'
#' @param input standard shiny parameter
#'
#' @param output standard shiny parameter
#'
#'@import magrittr
#'@import dplyr
#'@import shiny
#'@return None
server = function(input,output) {
options(shiny.maxRequestSize=300*1024^2) #increase the max file limit
#QUALITY CONTROL
rawData <- reactiveVal()
filteredData <- reactiveVal()
output$dgraph = renderPlot({
if(input$Help_QC == "Help") {
return(NULL)
}
#input raw data
rdata <- input$rdata
filePath <- input$rdata$datapath
fileType <- input$FileType
if (fileType == 'PBMC3k Test Data') {
inputData = LoadTestData()
} else {
inputData <- LoadData(fileType, filePath)
}
rawData(inputData)
#Data Processing
minNum = input$Min_Features
maxNum = input$Max_Features
percentMT = input$Mitochondria
nFeatures = input$nFeatures
cleanData <- FilterData(rawData(), minNum, maxNum, percentMT, nFeatures)
filteredData(cleanData)
if(input$DPOptions == "Quality Control") {
qcPlots <- VlnPlot(rawData(), features = c("nFeature_RNA", "nCount_RNA", "percent.mt"), ncol = 3)
return(qcPlots)
} else if (input$DPOptions == "Data Filtration") {
newQCplots <- VlnPlot(filteredData(), features = c("nFeature_RNA", "nCount_RNA", "percent.mt"), ncol = 3)
return(newQCplots)
}
})
#FEATURE SELECTION
featureGenes <- reactiveVal()# DATASET 2
featureList <- reactiveVal()
output$FSPlot = renderPlot({
if(is.null(filteredData()) || input$Help_FS == "Help") {
return(NULL)
}
fs <- input$FS
featureNumber <- input$nFeatures
fs_results <- featureSelection(filteredData(), fs, featureNumber)
featureGenes(fs_results$data)
featureList(fs_results$featurelist)
fs_plot <- fs_results$plot
return(fs_plot)
})
#DIMENSION REDUCTION
dimRData <- reactiveVal() # DATASET 3
dimRMethod <- reactiveVal()
output$drPlot = renderPlot({
if(is.null(featureGenes()) || input$Help_DR == "Help") {
return(NULL)
}
dimR <- input$dr
graphType_DR <- input$DRG_options
L <- input$L_value
if(dimR == "PCA (Seurat)") {
dr_data <- PCA_DimR(featureGenes(), graphType_DR)
dimRData(dr_data$data)
dimRMethod(dr_data$method)
return(dr_data$plot)
} else if(dimR == "GLM PCA") {
dr_data <- GLM_PCA_DimR(featureGenes(), L)
dimRData(dr_data$data)
dimRMethod(dr_data$method)
return(dr_data$plot)
} else if(dimR == "GLM PCA (Residuals)") {
dr_data <- GLM_PCA_Residuals_DimR(featureGenes(), graphType_DR)
dimRData(dr_data$data)
dimRMethod(dr_data$method)
return(dr_data$plot)
}
})
#Cluster Validation
output$cvPlot = renderPlot({
if(input$CVOptions == "Estimate K with SC3" || is.null(dimRData()) || input$Help_CV == "Help") {
return(NULL)
}
graphType_CV <- input$CVOptions
maxK <- input$maxK
cvData <- dimRData()
clusterValPlot <- clusterValidation(cvData, graphType_CV, maxK)
return(clusterValPlot)
})
#SC3 Estimate K
output$sc3_estimated_k_value <- renderText({
if(input$CVOptions != "Estimate K with SC3" || is.null(dimRData()) || input$Help_CV == "Help") {
return(NULL)
}
clusValData <- dimRData()
k <- sc3_estimate(clusValData)
string <- paste("Estimated number of clusters = ", k)
return(string)
})
#CLUSTER ANALYSIS
clusterData <- reactiveVal() # DATASET 4
output$SClusters = renderPlot({
if(is.null(dimRData()) || input$Help_CA == "Help") {
return(NULL)
}
sO <- dimRData() #this object should have both the current feature gene and reduced dimension sets already loaded.
reductionMethod <- dimRMethod() #formerly method1
clusteringMethod <- input$CM #formerly method
graphType_CA <- input$cv #formerly reductionMethod
noDim <- input$dimensions
res <- input$resolution
kValue <- input$k_value
coreNumber <- input$ncore
linkM <- input$LM
hcOption <- input$HCoptions
ha_fs_option <- input$DLoptions
ha_cl_option <- input$DLoptions2
if (clusteringMethod == "K-Nearest Neighbor (Seurat)") {
knn <- seuratClustering(sO, noDim, res, graphType_CA, reductionMethod)
clusterData(knn$data)
return(knn$plot)
} else if (clusteringMethod == "Graph Based Hierarchical Clustering (HGC)") {
hgc <- HGC_clustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(hgc$data)
return(hgc$plot)
} else if (clusteringMethod == "Hierarchical Clustering") {
hc <- h_clustering(sO, linkM, noDim, graphType_CA, reductionMethod, kValue, hcOption)
clusterData(hc$data)
return(hc$plot)
} else if (clusteringMethod == "K-means Clustering") {
km <- km_clustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(km$data)
return(km$plot)
} else if (clusteringMethod == "Density Peak Clustering (Monocle)") {
cd <- monocleClustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(cd$data)
return(cd$plot)
} else if (clusteringMethod == "Consensus Clustering (SC3)") {
cc <- consensus_clustering(sO, kValue, noDim, graphType_CA, reductionMethod)
clusterData(cc$data)
return(cc$plot)
}
})
#OUTPUT DATA
resultsTable = reactiveVal()
output$table = renderTable({
if(is.null(clusterData()) || input$Help_R == "Help") {
return(NULL)
}
tableChoice = input$TableOptions
sO = clusterData()
resultsTable = generateTable(sO, tableChoice)
return(resultsTable)
})
output$Pipeline <- renderUI({
str1 <- paste("<B>Feature Selection Algorithm:<B> ", input$FS)
str2 <- paste("<B>Dimension Reduction Method:<B> ", input$dr)
str3 <- paste("<B>Unsupervised Clustering Method:<B> ", input$CM)
HTML(paste(str1, str2, str3, sep = '<br/>'))
})
output$dlb <- downloadHandler(
filename <- "FinalData.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(resultsTable(), file, row.names = FALSE)
}
)
output$FSdlb <- downloadHandler(
filename <- "Selected_Features.txt",
contentType <- "csv",
content <- function(file) {
utils::write.csv(featureList(), file, row.names = FALSE)
}
)
output$SOdlb <- downloadHandler(
filename <- "FinalSeuratObject.txt",
contentType<- "RDS",
content <- function(file) {
saveRDS(clusterData(), file)
}
)
#Differentially Expressed Gene Analysis
output$degPlot = renderPlot({
if(is.null(clusterData()) || input$Help_DA == "Help") {
return(NULL)
}
sO <- clusterData()
lfc <- input$LFC
minPCT <- input$minPC
outputType <- input$DEG_Options
degResults = degAnalysis(sO, lfc, minPCT, outputType)
return(degResults)
})
degMarkerTable = reactiveVal()
output$degTable = renderTable({
if(is.null(clusterData()) || input$Help_DA == "Help") {
return(NULL)
}
sO <- clusterData()
lfc <- input$LFC
minPCT <- input$minPC
outputType <- input$DEG_Options
degResults = degAnalysis(sO, lfc, minPCT, outputType)
degMarkerTable(degResults)
return(degResults)
})
output$deg_biomarkers <- downloadHandler(
filename <- "ClusterBiomarkers.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(degMarkerTable, file, row.names = FALSE)
}
)
#Deep Learning
deepLearningData <- reactiveVal()
output$DL_Plot <- renderPlot({
if(is.null(dimRData()) || input$Help_DL == "Help") {
return(NULL)
}
#Variables
sO <- dimRData()
kValue <- input$DL_Knumber
noDim <- input$DL_dimensions
coreNumber <- input$DL_ncore
reductionMethod <- dimRMethod() #formerly method1
graphType_DL <- input$DL_cv #formerly reductionMethod
#DL Function Call
deepLearningResults <- autoEncoderClusterring(sO, noDim, kValue, coreNumber, graphType_DL, reductionMethod)
deepLearningData(deepLearningResults$data)
return(deepLearningResults$plot)
})
output$dl_seurat <- downloadHandler(
filename <- "DL_SeuratObject.txt",
contentType <- "RDS",
content <- function(file) {
saveRDS(deepLearningData(), file)
}
)
resultsTable_DL = reactiveVal()
output$DL_Table = renderTable({
if(is.null(deepLearningData()) || input$Help_DL == "Help") {
return(NULL)
}
sO = deepLearningData()
tableChoice = input$DL_TableOptions
if (tableChoice == "Summary Report") {
Dsummary <- table(Idents(sO))
Dsummary <- as.data.frame(Dsummary)
colnames(Dsummary) <- c("Cluster", "Frequency")
resultsTable_DL(Dsummary)
return(Dsummary)
} else {
results <- as.data.frame(sO@active.ident)
colnames(results) <- "Cluster"
results <- cbind(rownames(results), data.frame(results, row.names = NULL))
colnames(results) <- c("Cell Label", "Cluster")
results <- results[order(results$Cluster), ]
resultsTable_DL(results)
return(results)
}
})
output$dl_table_download <- downloadHandler(
filename <- "DL_Data.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(resultsTable_DL(), file, row.names = FALSE)
}
)
output$DL_Heatmap = renderPlot({
if(is.null(deepLearningData()) || input$Help_DL == "Help") {
return(NULL)
}
#Variables
sO <- deepLearningData()
lfc_DL <- input$DL_LFC
minPCT_DL <- input$DL_minPC
degHeatMap_DL = degAnalysis(sO, lfc_DL, minPCT_DL, "Heatmap")
#sO.markers_DL <- FindAllMarkers(sO, only.pos = TRUE, min.pct = minPCT_DL, logfc.threshold = lfc_DL)
#sO.markers_DL %>% group_by(cluster) %>% top_n(n = 2, wt = avg_log2FC) -> top10
#degHeatMap_DL <- DoHeatmap(sO, features = top10$gene) + NoLegend()
return(degHeatMap_DL)
})
biomarkerTable_DL = reactiveVal()
output$DL_Biomarkers = renderTable({
if(is.null(deepLearningData()) || input$Help_DL == "Help") {
return(NULL)
}
#Variables
sO <- deepLearningData()
lfc_DL <- input$DL_LFC
minPCT_DL <- input$DL_minPC
#sO.markers_DL <- FindAllMarkers(sO, only.pos = TRUE, min.pct = dl_minpct, logfc.threshold = dl_lfc)
#sO.markers_DL %>% group_by(cluster) %>% top_n(n = 2, wt = avg_log2FC) -> top10
#degTable_DL <- dl_sO.markers %>% group_by(cluster) %>% slice_max(n = 2, order_by = avg_log2FC)
degTable_DL = degAnalysis(sO, lfc_DL, minPCT_DL, "Table")
biomarkerTable_DL(degTable_DL)
return(degTable_DL)
})
output$dl_bmTable <- downloadHandler(
filename <- "DL_BiomarkerTable.csv",
contentType <- "csv",
content <- function(file) {
utils::write.csv(biomarkerTable_DL(), file, row.names = FALSE)
}
)
#The Help Text
output$QC_Help_Text <- renderText(
"This is the Data Processing Tab. Accepted inputs include Seurat/Single Cell Experiment Objects (as RDS files)
and Gene x Cell matrices (as CSV files). 'Quality Control' generates violin plots visualizing the QC metrics of the raw input data.
'Data Filtration' trims cells from the data according to user defined Feature Count and % Mitochondrial DNA limits.
Sample Test Data (PBMC3k) is also available here."
)
output$FS_Help_Text <- renderText(
"This is the Feature Selection Tab. Feature Selection identifies genes that show biologically revelent variation.
Here genes can be assessed and selected according to a variety of criteria, ranging from Variance to Drop-Out Rate."
)
output$DR_Help_Text <- renderText(
"This is the Dimension Reduction Tab. The purpose of dimension reduction is to compress input data into a 'lower dimensional' representation more compatible
with clustering. This can be achieved here using either PCA, GLM-PCA (a modification of PCA optimized for UMI count data), or Residual PCA
(a fast approximation of the GLM-PCA algorithm)."
)
output$CV_Help_Text <- renderText(
"This is the Cluster Validation Tab. This tab estimates the number of clusters in the data using either the Elbow or Silhouette Methods.
Cluster number can also be directly estimated using the Tracy-Widom Theory of Random Matrices (via SC3). This tab is both time-intensive and optional -
this part of the pipeline can safely be skipped."
)
output$CA_Help_Text <- renderText(
"This is the Cluster Analysis Tab."
)
output$R_Help_Text <- renderText(
"This is the R Help Text."
)
output$DA_Help_Text <- renderText(
"This is the Differentially Expressed Gene (DEG) Analysis Tab."
)
output$DL_Help_Text <- renderText(
"This is the Deep Learning Tab. This Tab represents a fully independent analysis pipeline."
)
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.