R/fun.R

In twesleyb/Uezu2016: iBioID Proteomics from Uezu, Kanak, Bradshaw et al., 2016

## ---- additional functions

normSL <- function(tp, groupBy=NULL){
#	suppressPackageStartupMessages({
#		library(dplyr)
#		library(data.table)
#	})
	# Create an expression that will be evaluated to dynamically 
	# group data based on user provided grouping factors. Default's
	# to Sample -- every replicate.
	tp <- ungroup(as.data.frame(tp))
	cmd <- paste0("group_by(tp, ",paste(groupBy,collapse=", "),")")
	# Calculate column sums for grouped samples.
	# FIXME: if .groups is set to 'drop' to avoid Warning msg, then error?
	data_list <- eval(parse(text=cmd)) %>% 
		summarize(Total=sum(Intensity,na.rm=TRUE),.groups="drop") %>%
		group_split()
	# Calculate normalization factors.
	data_SL <- lapply(data_list,function(x) {
				     x$"Mean(Total)" <- mean(x$Total,na.rm=TRUE)
				     x$NormFactor <- x$"Mean(Total)"/x$Total
				     x$Norm <- x$Total*x$NormFactor
			      return(x) }) %>% bind_rows()
	# Collect normalization factors as named vector.
	SL_factors <- data_SL$NormFactor
	names(SL_factors) <- as.character(data_SL[["Sample"]])
	# Normalize measurements by sample factors.
	tp$Intensity <- tp$Intensity * SL_factors[as.character(tp[["Sample"]])]
	return(as.data.table(tp))
} #EOF


normSP <- function(tp, pool){
	# perform normalization to pooled QC samples
	# store a copy of the data
	tp <- ungroup(tp)
	tp_copy <- tp
	# group pooled together
	tp$Group <- as.numeric(grepl(paste(pool,collapse="|"),tp$Sample))
	tp_list <- tp %>% group_by(Protein,Group) %>% 
		dplyr::summarize(Mean_Intensity=mean(Intensity,na.rm=TRUE),
			  n = length(Intensity), .groups="drop") %>%
	as.data.table() %>% 
	arrange(Protein,Group) %>% 
	group_by(Protein) %>% group_split()
	# loop to calculate normalization factors
        new_list <- list()
	for (i in 1:length(tp_list)){
		x <- tp_list[[i]]
		x$NormFactor <- c(x$Mean_Intensity[2]/x$Mean_Intensity[1],1)
		x$Norm_Mean_Intensity <- x$Mean_Intensity * x$NormFactor
		new_list[[i]] <- x
	}
	tp_list <- new_list
	# collect in a df
	df <- do.call(rbind,tp_list) %>% 
		dplyr::select(Protein,Group,NormFactor)
	# merge with input data
	tp_norm <- left_join(tp,df,by=c("Protein","Group"))
	# Perform normalization step.
	tp_norm$Intensity <- tp_norm$Intensity * tp_norm$NormFactor
	tp_norm <- tp_norm %>% dplyr::select(colnames(tp_copy))
	tp_norm <- as.data.table(tp_norm)
	# return the normalized data
	return(tp_norm)
} #EOF

normSP2 <- function(tp, pool){
	# perform normalization to pooled QC samples
	# store a copy of the data
	tp <- ungroup(tp)
	tp_copy <- tp
	# group pooled together
	tp$Group <- as.numeric(grepl(paste(pool,collapse="|"),tp$Condition))
	tp_list <- tp %>% group_by(Protein,Group) %>% 
		dplyr::summarize(Mean_Intensity=mean(Intensity,na.rm=TRUE),
			  n = length(Intensity), .groups="drop") %>%
	  as.data.table() %>% 
	  arrange(Protein,Group) %>% 
	  group_by(Protein) %>% group_split()
	# loop to calculate normalization factors
        new_list <- list()
	for (i in 1:length(tp_list)){
		x <- tp_list[[i]]
		x$NormFactor <- c(x$Mean_Intensity[2]/x$Mean_Intensity[1],1)
		x$Norm_Mean_Intensity <- x$Mean_Intensity * x$NormFactor
		new_list[[i]] <- x
	}
	tp_list <- new_list
	# collect in a df
	df <- do.call(rbind,tp_list) %>% 
		dplyr::select(Protein,Group,NormFactor)
	# merge with input data
	tp_norm <- left_join(tp,df,by=c("Protein","Group"))
	# Perform normalization step.
	tp_norm$Intensity <- tp_norm$Intensity * tp_norm$NormFactor
	tp_norm <- tp_norm %>% dplyr::select(colnames(tp_copy))
	tp_norm <- as.data.table(tp_norm)
	# return the normalized data
	return(tp_norm)
} #EOF



write_excel <- function(mydata, myfile, rowNames = FALSE, colNames = TRUE, ...) {
  # imports
  suppressPackageStartupMessages({
  })
  # check that input data is a list
  if ("list" %in% class(mydata)) {
    mylist <- mydata
  } else {
    # Coerce to list
    mylist <- list(mydata)
  }
  # Insure there are names.
  if (is.null(names(mylist))) {
    names(mylist) <- paste("Sheet", c(1:length(mylist)))
  }
  wb <- openxlsx::createWorkbook()
  # Loop to add a worksheets:
  for (i in 1:length(mylist)) {
    df <- as.data.frame(mylist[[i]])
    openxlsx::addWorksheet(wb, sheetName = names(mylist[i]))
    openxlsx::writeData(wb,
      sheet = i, df,
      rowNames = rowNames, colNames = colNames, ...
    )
  }
  # Save workbook.
  openxlsx::saveWorkbook(wb, myfile, overwrite = TRUE)
}