• Add 500 bp up and downstream of each probe

$ slopBed -i probes.bed -b 500 > probes.500bp.bed 

• Get a BED file of all regions not covered by the probes (+500 bp up/down)

$ complementBed -i probes.500bp.bed -g hg18.genome > probes.500bp.complement.bed 

• Create a masked genome where all bases are masked except for the probes +500bp

$ maskFastaFromBed -in hg18.fa -bed probes.500bp.complement.bed -fo hg18.probe-complement.masked.fa 

• Find all SNPs that are not in dbSnp and not in the latest 1000 genomes calls

$ intersectBed -a snp.calls.bed -b dbSnp.bed -v | intersectBed -a stdin -b 1KG.bed -v > snp.calls.novel.bed 

• By default, coverageBed counts any feature in A that overlaps B by >= 1 bp. If you want to require that a feature align entirely within B for it to be counted, you can first use intersectBed with the “-f 1.0” option.

$ intersectBed -a features.bed -b windows.bed -f 1.0 | coverageBed -a stdin -b windows.bed > windows.bed.coverage 

• The basics...

$ coverageBed -abam reads.bam -b exons.bed > exons.bed.coverage 

• Take it a step further and require that coverage be from properly-paired reads.

$ samtools view -uf 0x2 reads.bam | coverageBed -abam stdin -b exons.bed > exons.bed.proper.coverage 

• Use grep to only look at forward strand features (i.e. those that end in “+”).

$ bamToBed -i reads.bam | grep \+$ | coverageBed -a stdin -b genes.bed > genes.bed.forward.coverage 

• Use grep to only look at reverse strand features (i.e. those that end in “-”).

$ bamToBed -i reads.bam | grep \-$ | coverageBed -a stdin -b genes.bed > genes.bed.forward.coverage 

$ pairToPair -a sample1.sv.bedpe -b othersamples.sv.bedpe -type neither > sample1.sv.private.bedpe 

• We’ll require that 90% of the inner span of the deletion be overlapped by a recent ALU.

$ pairToBed -a deletions.sv.bedpe -b ALUs.recent.bed -type notispan -f 0.80 > deletions.notALUsinRef.bedpe