Missed a transform somewhere in Coereba_Concatenate

Author: DavidRach

NAMESPACE

@@ -1,6 +1,7 @@
 # Generated by roxygen2: do not edit by hand
 
 export(Coereba_App)
+export(Coereba_Concatenate)
 export(Coereba_FCSExport)
 export(Coereba_FCS_Reversal)
 export(Coereba_GateCutoffs)

R/Coereba_Concatenate.R

@@ -0,0 +1,46 @@
+#' Processes multiple Coereba flowframes, retrieves contents according to our Concatenate criteria, then cumulatively processes
+#' 
+#' @param Set The flow/cyto set we are working with
+#' @param metadata_cols The columns from pData to retain. 
+#' @param outpath Where to store the .fcs file
+#' @param filename What to call the .fcs file
+#' 
+#' @export
+#' 
+#' @examples A <- 2+2
+Coereba_Concatenate <- function(Set, metadata_cols, outpath=NULL, filename="CombinedCoerebaFile"){ 
+
+  AllData <- map(.x=Set, .f=Coereba_SingleFrame_Reversal, metadata_cols=metadata_cols) |> bind_rows()
+
+  if (is.null(outpath)){outpath <- getwd()}
+  
+  FCSFile <- Coereba_FCSExport(data=AllData, gs=Set[1],
+        returnType="fcs", outpath=outpath, filename=filename,
+        nameAppend="", Aggregate=FALSE, coerebaCombine =TRUE)
+}
+
+Coereba_SingleFrame_Reversal <- function(x, metadata_cols){
+
+  InternalPD <- pData(CoerebaGS)
+  TheMetadata <- InternalPD |> select(all_of(metadata_cols))
+  row.names(TheMetadata) <- NULL
+  CoerebaCS <- gs_pop_get_data(x)
+  CoerebaCF <- CoerebaCS[[1]]
+
+  if (class(CoerebaCF) %in% "cytoframe"){
+    Data <- exprs(CoerebaCF)
+    Data <- data.frame(Data, check.names=FALSE)
+    Original <- Data |> select(!starts_with("Coereba"))
+    Data <- Data |> select(starts_with("Coereba"))
+    Data <- Data |> mutate(across(everything(), as.character))
+    These <- colnames(Data)
+    Reverted <- map(.f=MetadataRetrieval, .x=These, data=Data, Coereba=CoerebaCF) |> bind_cols()
+    Assembled <- cbind(Original, Reverted)
+  
+    Completed <- Assembled |> cross_join(TheMetadata)
+    return(Completed)
+
+  } else {stop("Not a Cytoframe")}
+}
+
+

R/Coereba_FCS.R

@@ -7,10 +7,12 @@
 #' @param returnType Whether to return "flowframe" or "fcs"
 #' @param nameAppend For flowframe and fcs returnType, what gets appended before .fcs 
 #' @param Aggregate Combine the items, finish your documentation. 
+#' @param combineCoereba Internal, used for Coereba flowframe aggregation
 #'
 #' @importFrom flowWorkspace gs_pop_get_data keyword 
 #' @importFrom flowCore parameters exprs write.FCS parameters<-
-#' @importFrom dplyr select bind_cols
+#' @importFrom dplyr select bind_cols filter
+#' @importFrom stringr str_detect
 #' @importFrom tidyselect all_of
 #' @importFrom purrr map flatten
 #' @importFrom Biobase pData
@@ -22,13 +24,21 @@
 #' 
 #' @examples A <- 2+2
 Coereba_FCSExport <- function(data, gs, outpath, filename, fcsname,
-                              returnType, nameAppend, Aggregate=FALSE){
+                              returnType, nameAppend, Aggregate=FALSE, 
+                              coerebaCombine=FALSE){
   # Retrieving param information
   cs <- gs_pop_get_data(gs, "root")
   cf <- cs[[1]]
   original_param <- parameters(cf)
   original_descr <- keyword(cf)
   original_exprs <- exprs(cf)
+
+  if (coerebaCombine==TRUE){
+    paramdata <- original_param@data
+    paramdata <- paramdata |> filter(!str_detect(name, "Coereba"))
+    original_param@data <- paramdata
+    original_exprs <- original_exprs[, !grepl("Coereba", colnames(original_exprs))]
+  }
 
   # Identifying new columns
   OldColNames <- colnames(original_exprs) |> unname()

man/Coereba_Concatenate.Rd

@@ -591,6 +591,225 @@ RetrievedData <- Coereba::Coereba_FCS_Reversal(Coereba=FCSPath)
 head(RetrievedData, 5)
 ```
 
+## Normalization Bypass
+
+Alternatively, if you are planning to normalize the files with CytoNorm, you don't want to return a concatenated file, but individual cytoframe objects. Here is an example workflow:
+
+```{r}
+#| label: ListOfFlowFrames
+CoerebaFlowFrame <- Utility_Coereba(gs=UnmixedGatingSet, subsets="root",
+ sample.name="GROUPNAME", reference=CheckedGates, starter="Comp-PE-Vio770-A",
+ inverse.transform = TRUE, returnType="flowframe", Individual=TRUE, outpath=NULL,
+ columns = MarkersForToday)
+```
+
+```{r}
+CoerebaFlowFrame1 <- unlist(CoerebaFlowFrame)
+
+FlowSet <- flowCore::flowSet(CoerebaFlowFrame1)
+FlowGS <- GatingSet(FlowSet)
+
+MarkersUpdated <- gsub("Comp-", "", MarkersForToday)
+```
+
+```{r}
+MyBiexponentialTransform <- flowjo_biexp_trans(channelRange = 256, maxValue = 1000000,
+                                               pos = 4.5, neg = 0, widthBasis = -1000)
+TransformList <- transformerList(MarkersUpdated, MyBiexponentialTransform)
+
+InverseBiexponential <- flowjo_biexp(inverse=TRUE, channelRange=256, maxValue=1000000, pos=5.25, neg=0, width=-1000)
+ReverseBiexponential <- transformList(MarkersUpdated, InverseBiexponential)
+
+FlowSet_Transformed <- transform(FlowGS, TransformList)
+```
+
+```{r}
+ShinyGatesSplitpoints2 <- ShinyGatesSplitpoints
+colnames(ShinyGatesSplitpoints2) <- gsub("Comp-", "", colnames(ShinyGatesSplitpoints2))
+
+PostCoereba <- Utility_NbyNPlots(x=FlowGS[4], sample.name=c("GROUPNAME"), experiment="ABs",
+condition="DNs", removestrings=".fcs", marginsubset="root", gatesubset="root", 
+ycolumn="PE-Vio770-A", bins=150, clearance=0.1, 
+gatelines=TRUE, reference=ShinyGatesSplitpoints2, outpath=StorageLocationForPlotPDF, returntype="patchwork")
+```
+
+```{r}
+PostCoereba
+```
+
+### Readding filtering data
+
+```{r}
+Specimen <- function(x){
+  pattern <- "([A-Z]+[0-9]+_[0-9]+)(?=_SEB)"
+  result <- stringr::str_extract(x, pattern)
+  result <- data.frame(Specimen = result)
+  return(result)
+}
+
+Experiment <- function(x){
+  pattern <- "(?<=_)AB_[0-9]{2}(?=_)" 
+  result <- stringr::str_extract(x, pattern)
+  result <- data.frame(Experiment = result)
+  return(result)
+}
+```
+
+```{r}
+OriginalPD <- pData(UnmixedGatingSet)
+ValuesOriginal <- OriginalPD |> pull(name)
+TheSpecimen <- map(ValuesOriginal, .f=Specimen) %>% bind_rows()
+TheExperiment <- map(ValuesOriginal, .f=Experiment) %>% bind_rows()
+```
+
+```{r}
+OtherPD <- pData(FlowGS)
+UpdatedPD <- cbind(OtherPD, TheSpecimen, TheExperiment)
+UpdatedPD <- UpdatedPD |> mutate(Donor = str_replace(Specimen, "_[0-9]+$", ""))
+pData(FlowGS) <- UpdatedPD
+```
+
+### Subsetting Controls
+
+```{r}
+TheNormies <- subset(FlowGS, Donor=="NY068")
+#pData(TheNormies)
+NormCounts <- gs_pop_get_stats(TheNormies, nodes="root", type="count")
+TheCount <- sum(NormCounts$count)
+NormiesCS <- gs_cyto_data(TheNormies)
+#pData(NormiesCS)
+TheNewNormies <- cytoset_to_flowSet(NormiesCS)
+```
+
+```{r}
+Batch <- pData(TheNewNormies)$Experiment
+```
+
+```{r}
+library(CytoNorm)
+
+NormPath <- file.path("C:", "Users", "12692", "Desktop", "CytoNorm")
+
+model <- CytoNorm.train(files=TheNewNormies,
+                        labels=Batch,
+                        channels=MarkersUpdated,
+                        transformList=NULL,
+                        seed=1989,
+                        plot=TRUE,
+                        verbose=TRUE,
+                        normMethod.train = "QuantileNorm.train",
+                        normParams = list(nQ = 99),
+                        FlowSOM.params = list(nCells = TheCount,
+                          xdim = 15, ydim = 15, nClus = 9,
+                          scale = FALSE, colsToUse = MarkersUpdated),
+                        outputDir = NormPath,
+                        clean = TRUE,
+                        recompute = TRUE)
+```
+
+```{r}
+LongTermStorage <- file.path(NormPath, "SmallModel.rds")
+saveRDS(object = model, file = LongTermStorage)
+```
+
+```{r}
+LongTermStorage <- file.path(NormPath, "SmallModel.rds")
+previousModel <- readRDS(LongTermStorage)
+goal_q <- getCytoNormQuantiles(previousModel)
+Batch <- pData(TheNormies)$Experiment
+#ForClustering1 <- GetChannels(object = NormalizeGS2[[1]], markers = ForClustering)
+
+model_Distribution <- CytoNorm.train(files=TheNewNormies,
+                        labels=Batch,
+                        channels=MarkersUpdated,
+                        transformList=NULL,
+                        seed=1989,
+                        plot=TRUE,
+                        verbose=TRUE,
+                        normParams = list("goal" = goal_q),
+                        FlowSOM.params = list(nCells = TheCount,
+                          xdim = 15, ydim = 15, nClus = 9,
+                          scale = FALSE, colsToUse=MarkersUpdated),
+                        outputDir = NormPath,
+                        clean = TRUE,
+                        recompute = TRUE)
+```
+
+
+And finally, we normalize everyone else
+```{r}
+AllNormies <- FlowGS
+AllCS <- gs_cyto_data(AllNormies)
+AllFS <- cytoset_to_flowSet(AllCS)
+
+Batch <- pData(AllFS)$Experiment
+
+# Use model or model_Distribution
+
+NormalizeAttempt <- CytoNorm.normalize(model=model, 
+                                       files=AllFS,
+                                       labels=Batch,
+                                       transformList = NULL,
+                                       verbose=TRUE,
+                                       prefix="Norm_",
+                                       transformList.reverse = NULL,
+                                       outputDir= NormPath,
+                                       clean = TRUE, 
+                                       write = FALSE)
+```
+
+And finally, we update the metadata for the new GatingSet
+```{r}
+pd3 <- pData(NormalizeAttempt)
+colnames(pd3) <- "thename"
+
+pd <- pData(FlowGS)
+UpdatedPD <- cbind(pd3, pd)
+
+pData(NormalizeAttempt) <- UpdatedPD
+```
+
+```{r}
+#| eval: FALSE
+Population <- "TestSet"
+UpdatedGS <- GatingSet(NormalizeAttempt)
+pData(UpdatedGS)
+filename <- paste0(Population, "_Norm")
+
+Utility_RidgePlots(gs=UpdatedGS, subset="root", TheY="Specimen", TheFill="Experiment", outpath=NormPath, returntype="pdf", filename = filename, therows = 1)
+
+NormedND <- subset(UpdatedGS, Donor == "NY068")
+
+filename <- paste0(Population, "_Norm_ND006")
+
+Utility_RidgePlots(gs=NormedND, subset="root", TheY="Specimen", TheFill="Experiment", outpath=NormPath, returntype="pdf", filename = filename, therows = 1)
+```
+
+### Reverse Biexponential
+```{r}
+ReversedGS <- transform(NormalizeAttempt, ReverseBiexponential)
+ThisWork <- GatingSet(ReversedGS)
+```
+
+```{r}
+NormedNotND <- subset(ThisWork, Donor != "NY068")
+pData(NormedNotND)
+TheCalcs <- gs_pop_get_stats(NormedNotND, nodes="root", type="count")
+min(TheCalcs$count)
+# 1000
+```
+
+# Out to FCS File
+
+```{r}
+Goodbye <- Coereba_Concatenate(Set=NormedNotND, metadata_cols=c("Experiment", "Donor"), outpath="C:/Users/12692/Desktop", filename="DidItWork")
+```
+
+
+
+
+
+
 # Generating a Summarized Experiment File
 
 Taking the `Utility_Coereba()` output, we can combine it with metadata and panel information into a SummarizedExperiment object. This will permit us in turn to tie in to a lot of the existing Bioconductor project analysis infrastructure. We do this through the `Coereba_Processing()` function.