compute_counts_pca_diffbind.Rmd

---
title: "pca diffBind"
author: "Javier Rodriguez Hernaez"
output: flexdashboard::flex_dashboard
---

```{r setup, include=FALSE}
input_dir="/Users/javrodher/Work/RStudio-PRJs/PTCL-project_2023/data/hichip-chip_peaks/"
output_dir="/Users/javrodher/Work/RStudio-PRJs/PTCL-project_2023/results/"
peakome_file="/Users/javrodher/Work/RStudio-PRJs/PTCL-project_2023/data/peakome_tiled.bed"
ss_file="samples.diffBind.csv"
n_cores=8

options(scipen = 999)
knitr::opts_chunk$set(message=FALSE, warning=FALSE, echo=FALSE, results = FALSE)
knitr::opts_knit$set(root.dir = output_dir)
```

```{r}
### FUNCTIONS ###
FX1=function(sample_sheet,out_name,n_cores=1,peakome,stats_out=F){
  # sample_sheet=ss
  # out_name="all"
  # n_cores = n_cores
  # stats_out=T
      
  dba_data = dba(sampleSheet=sample_sheet,  minOverlap=0,filter=0, bRemoveM=F)
  print(dba_data)
  dba_data$config$cores = n_cores
  dba_data = dba.count(dba_data,peaks=peakome,
                     minOverlap=0,
                     bUseSummarizeOverlaps = F,
                     bParallel = T,
                     filter=0,
                     summits = 0)
  dba_data = dba.normalize(dba_data,normalize=DBA_NORM_LIB,bFullLibrarySize=TRUE)
  count_data=as.data.frame(dba_data$peaks)
  count_data=count_data[,c(1:3,grep("^RPKM",names(count_data))) ]
  names(count_data)[4:ncol(count_data)] = paste0(dba_data$samples$SampleID,".RPKM")
  count_data$coord=paste(count_data$Chr,count_data$Start,count_data$End,sep=":")
  if((sum(count_data$coord %in% peakome$coord)==nrow(count_data)) !=TRUE) { print("ERROR")}
  count_data=count_data[order(count_data$coord),]
  peakome=peakome[order(peakome$coord),]
  count_data$PEAK_ID=peakome$PEAK_ID
  count_data=count_data[,c(ncol(count_data),1:(ncol(count_data)-2))]
  write.csv(count_data,paste0(output_dir,"/counts_normLib_",out_name,".csv"),row.names = F)

  stats=dba.show(dba_data)
  print(stats)
  pdf(paste0(output_dir,"/pca_",out_name,".pdf"))
  dba.plotPCA(dba_data,attributes = DBA_TISSUE,label="ID",labelSize=0.5)
  dev.off()
  
  if(stats_out) {return(stats)}
}


```

```{r}
library(DiffBind)
ss = read.csv(paste0(input_dir,"/",ss_file))
ss[is.na(ss)]=0

# customize sample sheet
ss = ss[-c(10,13),]
ss$SampleID=gsub(x=ss$SampleID,pattern="-Arima-rep1|-untreated|-DMSO|-Naive|_NOS",replacement = "",fixed = F)

# load peakome
peakome=read.delim(peakome_file,stringsAsFactors = F,header = F)
peakome=peakome[,1:4]
names(peakome)=c("chr","start","end","PEAK_ID")
peakome$coord=paste(peakome$chr,peakome$start,peakome$end,sep = ":")
```

```{r}
stats=FX1(sample_sheet=ss,
      out_name="all",
      n_cores = n_cores,
      peakome,
      stats_out=T)

write.csv(stats,paste0(output_dir,"/stats_diffBind.csv"),row.names = F,quote=F)
```

```{r}
FX1(sample_sheet=ss[ss$SampleID!="IL2",],
      out_name="noIL2",
      n_cores = n_cores,
      peakome,
      stats_out=F)
```