---
title: "Data analysis of the biological replicates"
output:
  html_document: default
---

<br>

This notebook analyses the biological replicates and generates a box plot with the copy number for different clones.

<br>

- Select values for global variables:

    - *PLOIDY* indicates the ploidy of the cell line studied.
    - *ASSAY*: 
        - *dual_color*: two-fluorophores dPCR assay to assess specific or total number of integrations (first color) normalized against genomic DNA loading (second color).
        - *dual_color_all_tag_only*: dual-fluorophore dPCR assay to assess total number of tag integrations in the target genome (first color) normalized against the genomic DNA amount (second color)
        - *triple_color*: three-fluorophores based dPCR assay to assess at the same time specific and total number of integrations (first and second color) both normalized against the same DNA loading amount (third color).
    - *CONTROL_LABELS*: substrings of the sample names indicating when a sample is a control.

<br>

```{r}
PLOIDY <- 3

# ASSAY <- "dual_color"
# ASSAY <- "dual_color_all_tag_only"
# ASSAY <- "triple_color"
 ASSAY <- "triple_color_2023_screen"

CONTROL_LABELS <- c("U2OS", "Nup96", "Nup62", "NTC", "Comp", "wt", "HK-wt")
```

- Import functions.
```{r}
source("functions.R")
```

---

### Pre-processing data

---

- Read input files for each assay. For each *Sample*, there are *Positives* and *Negatives*. *TargetType* encodes the type of sample (*Reference* and *Unknown* for TARGETS).

```{r, echo=F}
if (ASSAY == "dual_color")
{
  repGFP.filename <- file.path(getwd(), "data", "biologicalReplicates", "190517_allGFP_HK_Nup93-mEGFP_PRNP.csv")
  repHDR.filename <- file.path(getwd(), "data", "biologicalReplicates", "190517_HDR_HK_Nup93-mEGFP_PRNP.csv")
  repGFP <- read.csv(repGFP.filename, header=T, stringsAsFactors=F)
  repHDR <- read.csv(repHDR.filename, header=T, stringsAsFactors=F) 
  repGFP$Marker <- "GFP"
  repHDR$Marker <- "HDR"
  dataREPLICATES <- rbind(repGFP, repHDR)
  rm(repGFP, repHDR)
}

if (ASSAY == "dual_color_all_tag_only")
{
  rep.filename <- file.path(getwd(), "data", "biologicalReplicates", "190917_allSNAP_U2OS_TPR-SNAP_PPP.csv")
  dataREPLICATES <- read.csv(rep.filename, header=T, stringsAsFactors=F)
  dataREPLICATES$Marker <- "SNAP"
}

if (ASSAY == "triple_color" | ASSAY=="triple_color_2023_screen")
{
  if (ASSAY == "triple_color")
  {
    rep.filename <- file.path(getwd(), "data", "biologicalReplicates", "200703_allGFP-HDR_U2OS-Nup93-mEGFP_PCRNP.csv")  
    inputData <- read.csv(rep.filename, header=T, stringsAsFactors=F)
  } 
  if (ASSAY == "triple_color_2023_screen")
  { 
    rep.filename <- file.path(getwd(), "data", "biologicalReplicates", "200618_triple_U2OS-PCRNP-Nup93-mEGFP_allClones.csv") 
    inputData <- read.csv(rep.filename, header=T, stringsAsFactors=F, sep=";")
    # Get rid of Separability score
    inputData <- inputData[!(colnames(inputData) %in% grep("SeparabilityScore",colnames(inputData), value=T))]
  }

  # Get rid of concentrations and ProcessedDilutionStock
  inputData <- inputData[!(colnames(inputData) %in% grep("Concentration",colnames(inputData), value=T))]
  inputData <- inputData[!(colnames(inputData) %in% grep("ProcessedDilutionStock",colnames(inputData), value=T))]
    
  columns_metadata <- c(1:6)
  columns_greenChannel <- grep("Green_Channel_", colnames(inputData))
  columns_blueChannel <- grep("Blue_Channel_", colnames(inputData))
  columns_redChannel <- grep("Red_Channel_", colnames(inputData))
  
  dataREPLICATES_green <- cbind(inputData[,columns_metadata], inputData[,columns_greenChannel]) 
  dataREPLICATES_green$Marker <- "GFP"
  dataREPLICATES_green$TargetType <- "Unknown"

  dataREPLICATES_blue <- cbind(inputData[,columns_metadata], inputData[,columns_blueChannel]) 
  dataREPLICATES_blue$Marker <- "HDR"
  dataREPLICATES_blue$TargetType <- "Unknown"
    
  dataREPLICATES_red_GFP <- cbind(inputData[,columns_metadata], inputData[,columns_redChannel]) 
  dataREPLICATES_red_GFP$Marker <- "GFP"
  dataREPLICATES_red_GFP$TargetType <- "Reference"
  
  dataREPLICATES_red_HDR <- cbind(inputData[,columns_metadata], inputData[,columns_redChannel]) 
  dataREPLICATES_red_HDR$Marker <- "HDR"
  dataREPLICATES_red_HDR$TargetType <- "Reference"
  
  # Make colnames homogeneous
  colnames(dataREPLICATES_green)[7] <- "SampleType"
  colnames(dataREPLICATES_green)[8] <- "Positives"
  colnames(dataREPLICATES_green)[9] <- "Negatives"
  
  colnames(dataREPLICATES_red_GFP)[7] <- "SampleType"
  colnames(dataREPLICATES_red_GFP)[8] <- "Positives"
  colnames(dataREPLICATES_red_GFP)[9] <- "Negatives"
  
  colnames(dataREPLICATES_blue)[7] <- "SampleType"
  colnames(dataREPLICATES_blue)[8] <- "Positives"
  colnames(dataREPLICATES_blue)[9] <- "Negatives"
  
  colnames(dataREPLICATES_red_HDR)[7] <- "SampleType"
  colnames(dataREPLICATES_red_HDR)[8] <- "Positives"
  colnames(dataREPLICATES_red_HDR)[9] <- "Negatives"  
  
  dataREPLICATES <- rbind(dataREPLICATES_green, dataREPLICATES_red_GFP, dataREPLICATES_blue, dataREPLICATES_red_HDR)
  colnames(dataREPLICATES)[3] <- "Sample"
}

if (ASSAY=="dual_color" | ASSAY=="dual_color_all_tag_only")
{
  dataREPLICATES <- convertToLongFormat(dataREPLICATES)
}

dataREPLICATES[dataREPLICATES$TargetType=="Unknown",]$TargetType <- "TARGET"
dataREPLICATES[dataREPLICATES$TargetType=="Reference",]$TargetType <- "REFERENCE"
```

- Remove control samples:

```{r}
samplesToDiscard <- unique(grep(paste(CONTROL_LABELS, collapse="|"), dataREPLICATES$Sample, value=TRUE))
dataREPLICATES  <- dataREPLICATES[!(dataREPLICATES$Sample %in% samplesToDiscard),]
```

- Remove samples with 0 positives:

```{r}
dataREPLICATES <- dataREPLICATES[dataREPLICATES$Positives!=0,]
```

- Extract **Group** from the **Sample** name:

```{r}
dataREPLICATES <- extractGroup(dataREPLICATES, ASSAY)
```

- Number of samples for each marker and target type:

```{r}
knitr::kable(table(dataREPLICATES[c("Marker","TargetType")]))
```

---

### Compute parameters

---


- Compute the **Negative Rate**:
```{r}
dataREPLICATES$NegativeRate <- computeNegativeRate(dataREPLICATES)
```

- Compute the **Ratio** and the **Confidence Interval**:
```{r}
dataREPLICATES <- computeConfidenceInterval(dataREPLICATES)
```

- Compute **Relative Error**:
```{r}
dataREPLICATES <- computeRelativeError(dataREPLICATES)
```

- Update the **Ratio** with the ploidy:
```{r}
dataREPLICATES$Ratio <- PLOIDY * dataREPLICATES$Ratio
```

---

### Quality Control

---


- Keep only samples with relative error < 20 %.
```{r}
dataREPLICATES <- dataREPLICATES[dataREPLICATES$RelErr < 0.2,]
```


---

### Box plot with the copy numbers by clone

---


```{r, message=F, warning=F, echo=F}
dataREPLICATES$Group <- as.factor(dataREPLICATES$Group)

library(ggplot2)
p <- ggplot(dataREPLICATES, aes(x=Group, y=Ratio, fill=Marker))
p <- p + geom_boxplot(outlier.shape=NA)
p <- p + geom_hline(yintercept=3, color="grey", size=1)
p <- p + geom_point(pch=19, position=position_jitterdodge(), size=1, color="red")
p <- p + scale_fill_manual(values=c("gray","white"), labels = c("total", "on-target"))
p <- p + xlab("Clone ID")
p <- p + ylab("Copy Number")
p <- p + theme_bw()
p <- p + theme(legend.title = element_blank(), 
               panel.grid.major = element_blank(), 
               panel.grid.minor = element_blank(),
               axis.title = element_text(size = 14))
p
```