#Information##

#The samples were labelled as follows:

#iTRAQ 1:

#Donor 1: 113 - Fresh, 114 - 24 hours, 115 - 72 hours and 116: 7 days ##113 used as reference in JH1.csv

#Donor 2: 117 - Fresh, 118 - 24 hours, 119 - 72 hours and 121: 7 days ##117 used as reference in JH2.csv

#iTRAQ 2: 

#Donor 3: 113 - Fresh, 114 - 24 hours, 115 - 72 hours and 116: 7 days ##113 used as reference in JH3.csv

#Donor 4: 117 - Fresh, 118 - 24 hours, 119 - 72 hours and 121: 7 days ##117 used as reference in JH4.csv

#iTRAQ 3:

#Donor 5: 116 - Fresh, 113 - 24 hours, 114 - 72 hours and 115: 7 days ##116 used as reference in JH5.csv

#The spreadsheets attached contain the original data for all the samples run in each iTRAQ 
#and therefore have both donors timepoints relative to only 1 of the donors fresh samples/spreadsheet. 
#File names are JH1.csv, Jh2.csv, JH3.csv, JH4.csv, and JH5.csv

## Script requires package limma#
#source("https://bioconductor.org/biocLite.R")
#biocLite("limma") 



##### Sort data files to extract relevant information #####

JH1<-read.csv(file="JH1.csv") ## reads in datafile JH1.csv
JH1$Accession <- matrix(unlist(strsplit(as.character(JH1$Accession),"|",fixed=T)),ncol=3,byrow=T)[,2] ##strips out simple accession number 

## make objects of relevant columns ##

unusedJH1<-JH1[,2]
peptidesJH1<-JH1[,8]
IDJH1<-JH1[,5:6]
JH1T24<-JH1[,9]
JH1T72<-JH1[,14]
JH1T168<-JH1[,19]
Donor1<-cbind(unusedJH1, peptidesJH1, IDJH1, JH1T24,JH1T72,JH1T168)	### Binds relvant columns into a single dataframe
dataJH1 <- Donor1[which(Donor1[,2]>1),]					### Subsets protiens identified with 2 or more peptides
dataJH1A <- dataJH1[which(Donor1[,1]>1),]				### unused>1 (95%)
dataJH1B <- dataJH1A[which(Donor1[,2]==1),]				### subsets proteins identified by 1 peptide
dataJH1C <- dataJH1B[which(dataJH1B[,1]>2),]				### subsets proteins identified by 1 peptide with 99% conifidence (unused >2)
JH1data<-rbind(dataJH1A,dataJH1C)					### binds rows of data with 2+ peptides with 95% or more confidence and single petide with 99% confidence 

##### Repeat data sorting for other data files###


###############################################################################################################

JH2<-read.csv(file="JH2.csv")
JH2$Accession <- matrix(unlist(strsplit(as.character(JH2$Accession),"|",fixed=T)),ncol=3,byrow=T)[,2]
unusedJH2<-JH2[,2]
peptidesJH2<-JH2[,8]
IDJH2<-JH2[,5:6]
JH2T24<-JH2[,9]
JH2T72<-JH2[,14]
JH2T168<-JH2[,19]
Donor2<-cbind(unusedJH2, peptidesJH2, IDJH2, JH2T24,JH2T72,JH2T168)
dataJH2 <- Donor2[which(Donor2[,2]>1),]		
dataJH2A <- dataJH2[which(Donor2[,1]>1),]	
dataJH2B <- dataJH2A[which(Donor2[,2]==1),]		
dataJH2C <- dataJH2B[which(dataJH2B[,1]>2),]		
JH2data<-rbind(dataJH2A,dataJH2C)


###############################################################################################################

JH3<-read.csv(file="JH3.csv")
JH3$Accession <- matrix(unlist(strsplit(as.character(JH3$Accession),"|",fixed=T)),ncol=3,byrow=T)[,2]
unusedJH3<-JH3[,2]
peptidesJH3<-JH3[,8]
IDJH3<-JH3[,5:6]
JH3T24<-JH3[,9]
JH3T72<-JH3[,14]
JH3T168<-JH3[,19]
Donor3<-cbind(unusedJH3, peptidesJH3, IDJH3, JH3T24, JH3T72, JH3T168)
dataJH3 <- Donor3[which(Donor3[,2]>1),]		
dataJH3A <- dataJH3[which(Donor3[,1]>1),]	
dataJH3B <- dataJH3A[which(Donor3[,2]==1),]		
dataJH3C <- dataJH3B[which(dataJH3B[,1]>2),]
JH3data<-rbind(dataJH3A,dataJH3C)

###############################################################################################################

JH4<-read.csv(file="JH4.csv")
JH4$Accession <- matrix(unlist(strsplit(as.character(JH4$Accession),"|",fixed=T)),ncol=3,byrow=T)[,2]
unusedJH4<-JH4[,2]
peptidesJH4<-JH4[,8]
IDJH4<-JH4[,5:6]
JH4T24<-JH4[,29]
JH4T72<-JH4[,34]
JH4T168<-JH4[,39]
Donor4<-cbind(unusedJH4, peptidesJH4, IDJH4, JH4T24, JH4T72, JH4T168)
dataJH4 <- Donor4[which(Donor4[,2]>1),]		
dataJH4A <- dataJH4[which(Donor4[,1]>1),]	
dataJH4B <- dataJH4A[which(Donor4[,2]==1),]		
dataJH4C <- dataJH4B[which(dataJH4B[,1]>2),]		
JH4data<-rbind(dataJH4A,dataJH4C)

###############################################################################################################

JH5<-read.csv(file="JH5.csv")
JH5$Accession <- matrix(unlist(strsplit(as.character(JH5$Accession),"|",fixed=T)),ncol=3,byrow=T)[,2]
unusedJH5<-JH5[,2]
peptidesJH5<-JH5[,8]
IDJH5<-JH5[,5:6]
JH5T24<-JH5[,9]
JH5T72<-JH5[,14]
JH5T168<-JH5[,19]
Donor5<-cbind(unusedJH5, peptidesJH5, IDJH5, JH5T24, JH5T72, JH5T168)
dataJH5 <- Donor5[which(Donor5[,2]>1),]		
dataJH5A <- dataJH5[which(Donor5[,1]>1),]	
dataJH5B <- dataJH5A[which(Donor5[,2]==1),]	
dataJH5C <- dataJH5B[which(dataJH5B[,1]>2),]		
JH5data<-rbind(dataJH5A,dataJH5C)

###############################################################################################################

### Extract list of unique protein IDs ###

AllIDs<-rbind(IDJH1, IDJH2, IDJH3, IDJH4, IDJH5)
uniqueIDs<-unique(AllIDs)
dim(uniqueIDs)							### returns number of unique IDs in dataset ###

###Extract required columns for each dataset###
Sorted1<-JH1data[c(3,5,6,7)]
Sorted1<-unique(Sorted1)
Sorted2<-JH2data[c(3,5,6,7)]
Sorted2<-unique(Sorted2)
Sorted3<-JH3data[c(3,5,6,7)]
Sorted3<-unique(Sorted3)
Sorted4<-JH4data[c(3,5,6,7)]
Sorted4<-unique(Sorted4)
Sorted5<-JH5data[c(3,5,6,7)]
Sorted5<-unique(Sorted5)


#MERGE DATA FILES###
Merge<-merge(uniqueIDs, Sorted1, ,by="Accession",incomparables=NA,all=TRUE)
Merge12<-merge(Merge,Sorted2,by="Accession",incomparables=NA,all=TRUE)
Merge123<-merge(Merge12,Sorted3,by="Accession",incomparables=NA,all=TRUE)
Merge1234<-merge(Merge123,Sorted4,by="Accession",incomparables=NA,all=TRUE)
AllMerged<-merge(Merge1234,Sorted5,by="Accession",incomparables=NA,all=TRUE)

##Clean data of duplicates caused by White spaces that where present in some files###
Accession<-AllMerged$Accession				##Separates Accession
Accession<-trimws(Accession)				##Removes white spaces
Name<-AllMerged$Name					##Spearates Name
Name<-trimws(Name)					##Removes white spaces
minusID<-AllMerged[c(3:17)]				##Removes ID wiht white spaces from dataframe
Trimeddata<-cbind(Accession, Name, minusID)		## Replaced ID data without white spaces
uniqueheatdata<-unique(Trimeddata)			## Removes duplicates

heatdata<-uniqueheatdata[complete.cases(uniqueheatdata), ]
heatmapdata<-as.matrix(heatdata[c(3:17)])		
tiff("JHdataheatmap.tiff",compression="lzw",width=180, height=180,units="mm",res=600)
par(oma=c(2,2,2,2))
heatmap(heatmapdata,labRow="")
dev.off()


###################################
uniquedata<-uniqueheatdata			##new object of unique and clean data
#### Add 1 values for 3 Fresh samples to eable stats####

uniquedata$Fresh1 <- rep(1,nrow(uniquedata)) 
uniquedata$Fresh2 <- rep(1,nrow(uniquedata)) 
uniquedata$Fresh3 <- rep(1,nrow(uniquedata)) 
  

data<-uniquedata[c(1,2,18:20,3:17)]
ctrldata<-data[c(3:5)]
data24h<-data[c(6,9,12,15,18)]
data72h<-data[c(7,10,13,16,19)]
data168h<-data[c(8,11,14,17,20)]
completedata<-data[complete.cases(data), ]

###limmadata###
################################################################################
### Statistical analysis of proteins quantified in every sample; Liner model ###
################################################################################

library(limma)
lm.data<-completedata
prots<-lm.data[,1:2]
lm.data<-lm.data[,-(1:2)]
lglm.data<-log2(lm.data)
treat<-as.factor(c("Control","Control","Control","D1","D3","D7","D1","D3","D7","D1","D3","D7","D1","D3","D7","D1","D3","D7"))
design<-model.matrix(~0+treat)
colnames(design)<-c("Control","D1", "D3", "D7")
fit<-lmFit(lglm.data,design)
contrasts<-makeContrasts("D1-Control","D3-Control", "D7-Control", levels=design)
fit2<-contrasts.fit(fit, contrasts)
rawt<- fit2$coef / fit2$stdev.unscaled / fit2$sigma
rawp<- 2 * pt( abs(rawt), df=fit2$df.residual, lower.tail=FALSE)

bh<-matrix(nrow=nrow(rawp), ncol=3)

bh[,1]<-p.adjust(rawp[,1], method="BH")
bh[,2]<-p.adjust(rawp[,2], method="BH")
bh[,3]<-p.adjust(rawp[,3], method="BH")

##subset data with +/- 20% change in quantity and rawp or bh ajusted p <0.05 #####
sub1<-which((rawp[,1]<0.05)&(fit2$coef[,1]>0.26|fit2$coef[,1]<(-0.26)))
sub1a<-which((bh[,1]<0.05)&(fit2$coef[,1]>0.26|fit2$coef[,1]<(-0.26)))
sub2<-which((rawp[,2]<0.05)&(fit2$coef[,2]>0.26|fit2$coef[,2]<(-0.26)))
sub2a<-which((bh[,2]<0.05)&(fit2$coef[,2]>0.26|fit2$coef[,2]<(-0.26)))
sub3<-which((rawp[,3]<0.05)&(fit2$coef[,3]>0.26|fit2$coef[,3]<(-0.26)))
sub3a<-which((bh[,3]<0.05)&(fit2$coef[,3]>0.26|fit2$coef[,3]<(-0.26)))

##Volcanoplots###
tiff("Control 24h volcano.tiff",compression="lzw",width=180, height=180,units="mm",res=600)
par(ps=14,cex=1.5,family="serif")
plot(fit2$coef[,1], rawp[,1], log="y",ylim=c(1,1e-10),xlim=c(-5,5),xlab="logFc", ylab="p val", main="1 Day")
points(fit2$coef[sub1,1], rawp[sub1,1], col="blue")
points(fit2$coef[sub1a,1], rawp[sub1a,1], col="red")
dev.off()

tiff("Control 72h volcano.tiff",compression="lzw",width=180, height=180,units="mm",res=600)
par(ps=14,cex=1.5,family="serif")
plot(fit2$coef[,2], rawp[,2], log="y",ylim=c(1,1e-10),xlim=c(-5,5),xlab="logFc", ylab="p val", main="3 Days")
points(fit2$coef[sub2,2], rawp[sub2,2], col="blue")
points(fit2$coef[sub2a,2], rawp[sub2a,2], col="red")
dev.off()

tiff("Control 7Day volcano.tiff",compression="lzw",width=180, height=180,units="mm",res=600)
par(ps=14,cex=1.5,family="serif")
plot(fit2$coef[,3], rawp[,3], log="y",ylim=c(1,1e-10),xlim=c(-5,5),xlab="logFc", ylab="p val", main="7 Days")
points(fit2$coef[sub3,3], rawp[sub3,3], col="blue")
points(fit2$coef[sub3a,3], rawp[sub3a,3], col="red")
dev.off()

output<-cbind(prots,lm.data,fit2$coef,rawp,bh)
write.csv(output,file="JHstatistics.csv")
###################