peaks-report.Rmd

---
title: "Peaks per Sample Report"
author: "`r system('whoami', intern = TRUE)`"
date: "`r format(Sys.time(), '%B %d, %Y')`"
output:
  html_document:
    keep_md: true
    df_print: paged
params:
   input_dir: "/path/to/peaks_dir"
   plot_height: 12
   plot_width: 10
   report_name: "peaks_report"
---

```{r setup, include=FALSE}
# ~~~~~ CHUNK OPTIONS ~~~~~ # 
knitr::opts_chunk$set(echo = FALSE, warning = FALSE)

# ~~~~~ LIBRARIES ~~~~~ # 
library("ggplot2")
library("DT")
library("scales")
source("peaktools.R")

# ~~~~~ FUNCTIONS ~~~~~ # 
find_all_beds <- function (input_dirs, name_pattern = FALSE) {
    # find all .bed files in the supplied dirs
    bed_files <- dir(input_dirs, pattern = '.bed', full.names = TRUE, recursive = TRUE)
    if(name_pattern != FALSE) bed_files <- bed_files[which(basename(bed_files) %in% as.character(name_pattern))]
    return(bed_files)
}

get_numlines <- function(input_file) {
    # count the number of lines in a file
    return(length(readLines(input_file)))
}

mycat <- function(text){
    # function for formatting text in the report
    cat(gsub(pattern = "\n", replacement = "  \n", x = text))
}

split_sampleID_col <- function(df, new_colnames = NA, sample_colname = "sample", split_char = "-"){
    # split the 'sample' column in the df into separate columns, return the new df with all columns
    split_df <- as.data.frame(do.call(rbind, strsplit(x = df[[sample_colname]], split = split_char)))
    if (! is.na(new_colnames)) colnames(split_df) <- new_colnames
    df <- cbind(split_df, df)
    return(df)
}

make_bed_df <- function(input_dir){
    # create a dataframe from the bed files found in the dir
    bed_files <- find_all_beds(input_dirs = input_dir, name_pattern = "peaks.bed") 
    
    # sample ID is saved in the dirname of each bed file
    samples <- basename(dirname(bed_files))
    
    bed_df <- data.frame(sample = basename(dirname(bed_files)), 
                         peaks = sapply(X = bed_files, FUN = get_numlines), 
                         # file = bed_files, 
                         stringsAsFactors = FALSE)
    
    # split the sample ID's into separate columns
    bed_df <- split_sampleID_col(df = bed_df, new_colnames = c("patient", "status", "mark")) # ABC-R-H3K9AC
    
    # reindex the rownames
    rownames(bed_df) <- seq(nrow(bed_df))
    return(bed_df)
}

make_peaks_barplot <- function(df){
    # make the ggplot bar plot
    
    # re order the df
    # bed_df[order(bed_df[["peaks"]], decreasing = TRUE), ]
    
    peaks_barplot <- ggplot(data = df, aes(x = sample, y = peaks)) + geom_bar(stat = 'identity') + coord_flip() + ggtitle("Peaks per Sample") + scale_y_continuous(labels = comma)
    
    return(peaks_barplot)
}
```

```{r load_data, cache=TRUE, include=FALSE}
# ~~~~~ GLOBALS ~~~~~ #
input_dir <- normalizePath(params$input_dir)
plot_height <- as.numeric(params$plot_height)
plot_width <- as.numeric(params$plot_width)
report_name <- as.character(params$report_name)

bed_df <- make_bed_df(input_dir = input_dir)

peaks_barplot <- make_peaks_barplot(df = bed_df)

save.image(file = "loaded_data.Rdata", compress = TRUE)
write.table(x = bed_df, file = sprintf("%s.csv", report_name), quote = FALSE, sep = ',', row.names = FALSE, col.names = TRUE)
```

# Input dir

```{r}
mycat(input_dir)
```

# Peaks per Sample

## Plot

```{r, fig.height = plot_height, fig.width = plot_width}
peaks_barplot
```

## Table

```{r}
datatable(data = bed_df, rownames = FALSE, options = list(pageLength = -1))
```