LILY is a pipeline for detection of super-enhancers using H3K27ac ChIP-seq data, which includes explicit correction for copy number variation inherent to cancer samples. The pipeline is based on the ROSE algorithm originally developed by the the Young lab.
Follow steps 1-3 provided in the LILY's github documentation (https://github.com/BoevaLab/LILY). Clone the repository to run LILY scripts.
You will need following files to run LILY:
- narrowPeak, regions.bed and .wig files for a particular sample should be present in the data folder
- hg19_refseq.ucsc (transcriptome information which can be found at https://github.com/linlabbcm/rose2/tree/master/rose2/annotation)
- hg19.chrom.sizes (file with chromosome lengths)
Script calls super enhances iteratively for all lines. Make sure to change data directory and result directory paths before running the script.
Rscript lily.RAnnotating LILY calls using Homer
Change paths to directories to reflect paths to your files.
annotatePeaks.pl lilyCalls.bed hg19 > lillyCalls.anno.bedChipSeq heatmaps were generated for MYCN (annotating top 5K peaks) and all histone marks for COGN415 line (annotating filtered SEs from step 2)
To generate MYCN heatmaps (with top 5K peaks)
Rscript makeHeatmaps.RTo generate COGN415-histone-marks (with filtered SEs)
computeMatrix reference-point -S bigwigs/COGN415-H3K27Ac.bw bigwigs/COGN415-H3K27me3.bw bigwigs/COGN415-H3K4me1.bw bigwigs/COGN415-H3K4me3.bw -R SE_filtered/SE_filtered.bed -a 4000 -b 4000 --sortUsing max --skipZeros -o COGN415.mat.gz
plotHeatmap -m COGN415.mat.gz --colorList 'white,#ff7400' 'white,#004000' 'white,#00007F' 'white,#6F326F' --whatToShow 'heatmap and colorbar' --sortRegions descend --sortUsing max --zMin 0 --zMax 4 --regionsLabel Super_Enhancers -out COGN415.png
plotProfile -m COGN415.mat.gz -out COGN415.profile.png --perGroup --colors orange green blue purple --regionsLabel Super_Enhancers --plotHeight 10 --plotWidth 15
Author: patelk26@email.chop.edu
Organization: The Children's Hospital of Philadelphia (CHOP)