Report_CEL.Rmd

---
title: "Pipeline CEL Report"
author: "Sergi Aguiló Castillo"
date: "`r date()`"
output:
  html_document:
    toc: true
    toc_float: true
params: 
  Folder: ""
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r echo=FALSE, message=FALSE, results='hide'}

###Import libraries
library(ggplot2)
library(affyPLM)
library(affy)

#### Read the files
celfiles <- affy::ReadAffy(celfile.path = params$Folder)

```

# Pseudo-images

```{r, echo=FALSE, cache=FALSE, results=FALSE, warning=FALSE,  comment=FALSE, message= FALSE, eval =T, out.width='.7\\linewidth', fig.asp=1, fig.ncol = 1}

### Change the class of the files with a robut linear model
### Necessary for creating the Pseudo-images
cel_PLM <- affyPLM::fitPLM(celfiles)


# Make the image for each file
for (i in celfiles$sample){
  # Weighted image
  image(cel_PLM, which=i, add.legend = T)
  # Residuals
  image(cel_PLM, which=i, type="resids", add.legend = T)
}

```

# Processed data

```{r, echo=FALSE, results='hide'}

### Apply RMA and extract the expression values
norm_cel <- rma(celfiles)
norm_exp <- exprs(norm_cel)

# Vectors to do the dataframe
norm_sampleNames <- rep(colnames(norm_exp),each = nrow(norm_exp))
norm_logs = vector()
for (i in celfiles$sample){
  norm_logs = c(norm_logs, norm_exp[,i])
}
# Create a data frame
norm_logData = data.frame(logInt=norm_logs,sampleName=norm_sampleNames)

```

### Boxplot

```{r, echo=FALSE, results='hide'}

### Change size of the plot
par(mar=c(20,8,4.1,2.1))

ggplot(norm_logData, aes(sampleName, logInt)) + geom_boxplot() +
  xlab("Sample number") + ylab("Normalised expression values") +
  scale_y_log10() + 
  theme(axis.text.x = element_text(colour = "aquamarine4",
                                   angle = 50, size = 6.5, hjust = 1 ,
                                   face = "bold"))
```  

### Density plot

```{r, echo=FALSE, results='hide'}

ggplot(norm_logData, aes(logInt, color= sampleName)) +
  geom_line(stat="density") + scale_x_log10() +
  scale_color_discrete("Sample Name") + xlab("Normalised expression values")

```

# Comparison plots
### MA plot

```{r,echo=FALSE, fig.align='center'}

for (i in celfiles$sample){
  affy::MAplot(norm_cel, which = i)
}

```

### Relative Log Expression (RLE)

```{r, echo=FALSE, fig.width=10, fig.height=9}

### RLE can also be done by the following command
# RLE(cel_PLM, names = celfiles$sample)

par(mar = c(16, 10, 4.1, 2.1))
Mbox(norm_cel, main = "RLE", ylim = c(-0.4, 0.4),
     outline = FALSE, col = "mistyrose",
     las = 3, whisklty = 0, staplelty = 0, ylab = "RLE value")
abline(h=0)

```

### Normalised Unscaled Standard Error

```{r, echo=FALSE,fig.width=10, fig.height=9}

par(mar = c(16, 10, 4.1, 2.1))
boxplot(cel_PLM, main = "NUSE", outline = F,
        ylim = c(0.95,1.15),col = "lightblue", las = 3,
        whisklty = 0, staplelty = 0, ylab = "NUSE value")
abline(h=1)

```