unionbedg_with_NAs_context.sh

#!/bin/sh
#author: Cristian Peña
#Date: 19/06/2020
#Description: This script looks into the WGBS output folder (recomendation: first filter it by coverage) and uses a list of samples (from a text file, first column)
#             to extract the correspondent bedGraphs and context. Then it uses bedtools unionbedg to merge all samples (only the methylation column -4th column),
#             and finally, it filters every position based in a fraction of allowed NAs (eg. 0.2 = maximum 20% of NAs per position)
#             Difference with unionbedg_with_NAs.sh , this scripts allows to run each context in separate screens (faster)

#Usage: copy this script in a new empty directory where you want to create the new output. Also copy the sample text file in this directory. Create a new screen and run the script.
#bash unionbedg_with_NAs_context.sh {absolute_path_to_WGBS_output_folder} {samples.txt} {output_directory_name} {allowed_NAs_fraction} {context}

#example:
#bash unionbedg_with_NAs_context.sh /mnt/nfs/bioinfdata/home/NIOO/cristianp/WGBS/filtered samplesDMR_ALL.txt output_ALL_with_NAs 0.2 CpG         #in screen A
#bash unionbedg_with_NAs_context.sh /mnt/nfs/bioinfdata/home/NIOO/cristianp/WGBS/filtered samplesDMR_ALL.txt output_ALL_with_NAs 0.2 CHG         #in screen B
#bash unionbedg_with_NAs_context.sh /mnt/nfs/bioinfdata/home/NIOO/cristianp/WGBS/filtered samplesDMR_ALL.txt output_ALL_with_NAs 0.2 CHH         #in screen C

inputDir=$1
samples=($(cut -f1 ${2} | tr "\n" " "))
outputDir=$3
percent=$4
context=$5
number_samples=${#samples[@]}
allowed_NAs=$(printf "%.0f" $(echo ${number_samples}*${percent} | bc -l))
mkdir ${outputDir}

echo "processing "$number_samples" samples in "$context
echo "NAs allowed: "$allowed_NAs" ("$4")"
echo "process: extracting and reformatting samples"
for i in "${samples[@]}"
do
 id=${inputDir}/${i}/bedGraph/${i}_${context}.bedGraph
 echo ${outputDir}/${i}_${context}.txt >> ${outputDir}/samples_${context}.txt
 cut -f-4 ${id} > ${outputDir}/${i}_${context}.txt
done
input_bedtools=$(cat ${outputDir}/samples_${context}.txt | tr "\n" " ")
echo "process: bedtools unionbedg"
bedtools unionbedg -i ${input_bedtools} -filler NA -header -names ${samples[@]} > ${outputDir}/nonfiltered_${context}.bed
echo "process: filtering positions with more than "$allowed_NAs" NAs"
cat ${outputDir}/nonfiltered_${context}.bed | awk -v var="$allowed_NAs" '{count=0;for (i=4;i<=NF;i++){ if ($i=="NA"){count=count+1}}}{if (count<=var) print $0}' > ${outputDir}/unsorted_${context}.bed
echo "process: sorting bed file"
head -1 ${outputDir}/unsorted_${context}.bed > ${outputDir}/header_${context}
tail -n+2 ${outputDir}/unsorted_${context}.bed | sort -k1,1 -k2,2n > ${outputDir}/tmp_${context}.bed
cat ${outputDir}/header_${context} ${outputDir}/tmp_${context}.bed > ${outputDir}/${context}.bed
all_positions=$(wc -l ${outputDir}/nonfiltered_${context}.bed)
retained=$(wc -l ${outputDir}/${context}.bed)
echo "positions passing NA filter: "$retained"   from a total of "$all_positions
echo "process: removing temporary files"
rm ${outputDir}/*_${context}.txt
rm ${outputDir}/nonfiltered_${context}.bed
rm ${outputDir}/unsorted_${context}.bed
rm ${outputDir}/header_${context}
rm ${outputDir}/tmp_${context}.bed