Optimal BED file

In order to maximize the usage of RRBS data, it is recommendable to generate a taylored BED file to pinpoint the regions with data to be analyzed by CpelTdm. The following command allows to calculate the coverage of the i-th BAM file.

bedtools genomecov -ibam bam_file_i -bg > cov_i.bed

Regions in each cov_i.bed can be filtered on the basis of number of overlapping reads. For instance, the following command allows to retain those regions overlapping with more than 3 reads:

cat cov_i.bed | awk '{if($4>3){print $0}}' > cov_i.bed.tmp
mv cov_i.bed.tmp cov_i.bed

Next, regions that are within a given distance can be merged. For example, the following command allows to merge regions that are within 50 bp:

bedtools merge -d 50 -i cov_i.bed > cov_i.bed.merged
mv cov_i.bed.merged cov_i.bed

Finally, one might want to intersect or take the union of the regions obtained for multiple samples. The following commands create a file containing the union of the regions across samples:

cat cov_{1..N}.bed > cov_union.bed.tmp
sort-bed cov_union.bed.tmp > cov_union.bed
bedtools merge -d 50 -i cov_union.bed > cov_union.merged.bed
rm cov_union.bed.tmp cov_union.bed