From 788834dd790ca6d024f704be6cb685756857a503 Mon Sep 17 00:00:00 2001
From: Efrain Gonzalez <egonz340@fiu.edu>
Date: Mon, 12 Jun 2017 13:18:49 -0400
Subject: [PATCH] This code takes the clean data and discretizes it

---
 RPostClean.R | 157 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 157 insertions(+)
 create mode 100644 RPostClean.R

diff --git a/RPostClean.R b/RPostClean.R
new file mode 100644
index 0000000..ecb7946
--- /dev/null
+++ b/RPostClean.R
@@ -0,0 +1,157 @@
+#For Reading Raw Data from the created file
+
+#Required Libraries
+library(MASS)
+library(dplyr)
+library(tidyr)
+library(readr)
+library(stringr)
+
+
+#Necessary Functions
+
+#1# Function for discretizing the data 
+dndat <- function(NDATA){
+	rownd <- dim(NDATA)[1]
+	colnd <- dim(NDATA)[2]
+	DDATA <- matrix(0,nrow=rownd,ncol=colnd)
+	colnames(DDATA) <- colnames(NDATA)
+	i = 1
+	for(i in 1:rownd){
+		for(j in 1:colnd){
+			if(is.na(NDATA[i,j])==FALSE){
+			
+				if(NDATA[i,j] < -1){
+					DDATA[i,j]=0L
+				}
+				if(NDATA[i,j] > 1){
+					DDATA[i,j]=2L
+				}
+				if(-1 <= NDATA[i,j] && NDATA[i,j] < 1){
+					DDATA[i,j]=1L
+				}
+			} else{
+				DDATA[i,j] = NDATA[i,j]
+			}
+			j = j + 1			
+		}
+		i = i + 1
+	}
+	DDATA
+}
+
+
+#Bringing in the file
+rawdat <- file.choose()
+RAWDAT <- rawdat %>% 
+	read_delim(delim ="\t",col_names = FALSE,skip=1) %>%
+	filter(.,!grepl("Group|Age|Region|PMI|Title|Sex|Braak",X1))
+attributes(RAWDAT)$names <- RAWDAT[1,]
+
+#Just the clinical data
+RAWWORD <- rawdat %>%
+	read_delim(delim ="\t",col_names = FALSE,skip=1) %>%
+	filter(.,grepl("Group|Age|Region|PMI|Title|Sex|Braak",X1))
+attributes(RAWWORD)$names <- RAWDAT[1,]
+#Add col of NAs to clinical data
+z <- 1
+naroww <- as.data.frame(rep(0,dim(RAWWORD)[1]),stringsAsFactors = FALSE)
+for(z in 1:dim(RAWWORD)[1]){
+	naroww[z,1] <- as.integer(sum(is.na(RAWWORD[z,])))
+	z <- z + 1
+}
+colnames(naroww) <- "ROW_NAs"
+RAWWORD <- bind_cols(RAWWORD,naroww)
+
+
+##Getting back to the data
+RAWDAT2 <- RAWDAT[-1,] %>%
+	dplyr::arrange(.,ID_REF)
+
+##Editing the file for R processing
+RAWDATID <- RAWDAT2[,1] %>%
+	as.matrix(.)
+RAWDATNUM <- RAWDAT2[,-1] %>%
+	mapply(.,FUN = as.numeric) %>%
+	t(.)
+
+##Consolidating genes with the same name
+tabRDATID <- table(RAWDATID)
+NuRDATN <- matrix(0, nrow = dim(RAWDATNUM)[1], ncol = length(tabRDATID))
+j <- 1
+for(j in 1:length(tabRDATID)){
+	##Putting the ones without duplicates in their new homes
+	if(tabRDATID[j] == 1){
+		NuRDATN[,j] <- RAWDATNUM[,which(RAWDATID==rownames(tabRDATID)[j])]
+	}
+	##Averaging duplicates and putting them in their new homes
+	if(tabRDATID[j] > 1){
+		NuRDATN[,j] <- rowMeans(RAWDATNUM[,which(RAWDATID==rownames(tabRDATID)[j])],na.rm = TRUE)
+	}
+	j <- j + 1
+}
+
+
+#Scaling the Data
+scrawdat <- NuRDATN%>%
+	scale()
+attr(scrawdat,"scaled:center") <- NULL
+attr(scrawdat,"scaled:scale") <- NULL
+colnames(scrawdat) <- rownames(tabRDATID)
+
+
+#Discretized the Data
+dialzdat <- scrawdat %>%
+	dndat(.) %>%
+	t()%>%
+	as.data.frame(.)
+colnames(dialzdat) <- rownames(RAWDATNUM)
+
+#gene names
+genena <- as.data.frame(as.matrix(rownames(dialzdat),ncol=1))
+#setting "ID_REF" as a new variable
+colnames(genena) <- "ID_REF"
+rownames(dialzdat) <- NULL
+dialzdat <-bind_cols(genena,dialzdat)
+
+#NAs in a column
+x <- 2
+nacol <- as.data.frame(t(rep(0,dim(dialzdat)[2])),stringsAsFactors = FALSE)
+nacol[1,1] = "COL_NAs"
+for(x in 2:dim(dialzdat)[2]){
+	nacol[1,x] <- as.integer(sum(is.na(dialzdat[,x])))
+	x <- x + 1
+}
+colnames(nacol) <- colnames(dialzdat)
+dialzdat<-bind_rows(dialzdat,nacol)
+
+#NAs in a row
+y <- 1
+narowd <- as.data.frame(rep(0,dim(dialzdat)[1]),stringsAsFactors = FALSE)
+for(y in 1:dim(dialzdat)[1]){
+	narowd[y,1] <- as.integer(sum(is.na(dialzdat[y,])))
+	y <- y + 1
+}
+colnames(narowd) <- "ROW_NAs"
+dialzdat <- bind_cols(dialzdat,narowd)
+
+#converting to character so that the clinical can be brought together with discrete data
+k <- 2
+for(k in 2:dim(dialzdat)[2]-1){
+	dialzdat[,k] <- as.character(dialzdat[,k])
+	k <- k + 1
+}
+
+
+#The End the full data we seem to have found Carmen
+Fullalzdw <- bind_rows(RAWWORD,dialzdat)
+
+#Create the file
+nfnaex <- strsplit(rawdat,"[\\|/]") %>%
+	.[[1]] %>%
+	.[length(.)] %>%
+	gsub("\\D","",.) %>%
+	c("GSE",.,"dscrt.txt") %>%
+	paste(collapse = "")
+write.table(Fullalzdw, file = nfnaex, sep = "\t",col.names = TRUE,row.names = FALSE)
+
-- 
2.29.0