Compared to re-alignment approaches, liftover is a more rapid and cost-effective solution. To further increase the accuracy of liftover epigenome (WGBS and ChIP-Seq) and avoid misleading results, we implemented a three-step procedure to remove problematic regions (Figure 1) and ensure a more robust conversion between reference genome assemblies, namely, Lifted.
Lifted is part of our study which published on NAR Genomics and Bioinformatics. If you have used Lifted in a scientific publication, we would appreciate citations to the following paper:
Luu PL, Ong PT, Dinh TP, Clark SJ. Benchmark study comparing liftover tools for genome conversion of epigenome sequencing data. NAR Genomics and Bioinformatics. 2020 Sep;2(3):lqaa054.
Lifted gets inputs (BED files containing coordinates in hg19) and chain file (hg19ToHg38 from UCSC Genome Browser/NCBI) to generate BED files containing coordinates in hg38.
- First, Lifted removes all the gapped regions that cause corruption, including gapped-in-hg19, gapped-in-both (these regions are explained in Figure 1) and blacklist with bedtools.
- Second, the remaining coordinates are all ungapped on hg19 and then are ready to be converted by UCSC liftOver.
- Third, the inappropriate data such as duplication, alternative chromosome and not CG are removed. All output coordinates that overlap with gapped-in-hg38 are also removed by bedtools.
- First, Lifted removes all the gapped regions that cause corruption, including gapped-in-hg19, gapped-in-both, gapped-in-hg38 and blacklist with bedtools. For gapped-in-hg38, the input intervals in hg19 that overlap the coordinates of gapped-in-hg38 are split before liftover to cut out 2bp as presented in Figure 2.
- Second, the remaining coordinates are all ungapped on hg19 and then are ready to be converted by UCSC liftOver.
- Third, the inappropriate data such as duplication, alternative chromosome and not CG are removed by bedtools.
- First, Lifted removes all the gapped regions that cause corruption, including gapped-in-hg19, gapped-in-both (these regions are explained in Figure 1) and blacklist with bedtools.
- Second, the remaining coordinates are all ungapped on hg19 and then are ready to be converted by UCSC liftOver.
- Third, the inappropriate data such as duplication and alternative chromosome are removed. All output coordinates that overlap with gapped-in-hg38 are also removed by bedtools.
Figure 2. Two scenarios of liftover with cLifted (1) interval does not cross gapped-in-hg38, (2) interval crosses gapped-in-hg38.
- First, Lifted removes all the gapped regions that cause corruption, including gapped-in-hg19, gapped-in-both, gapped-in-hg38 and blacklist with bedtools. For gapped-in-hg38, the input intervals in hg19 that overlap the coordinates of gapped-in-hg38 are split before liftover to cut out 2bp as presented in Figure 2.
- Second, the remaining coordinates are all ungapped on hg19 and then are ready to be converted by UCSC liftOver.
- Third, the inappropriate data such as duplication and alternative chromosome are removed by bedtools.
Figure 4. Two scenarios of liftover with lLifted (1) interval does not cross gapped-in-hg38, (2) interval crosses gapped-in-hg38.
- Input BED files must have format: chromosome, start, end.
- Download chain file http://hgdownload.soe.ucsc.edu/goldenPath/hg19/liftOver/
- Download and install UCSC liftOver from http://hgdownload.soe.ucsc.edu/admin/exe/. UCSC liftOver needs to be exported: export PATH=$PATH:/path/to/file/liftOvertool/
- Download and install bedtools from https://bedtools.readthedocs.io/en/latest/
- Coordinates of filtered regions/positions are available at data folder (gapped-in-hg19.bed, gapped-in-both.bed, gapped-in-hg38.bed, duplication.bed, blacklist.hg19.bed and notCG.bed).
- Clone our Lifted repository using: git clone https://github.com/phuoc362/Lifted.git
Make sure the working directory is Lifted/bin
cd Lifted/bin
sh cLifted_CpG.sh <chain_file> <input.bed> <output.lifted.bed> <output.unlifted.bed>
sh lLifted_CpG.sh <chain_file> <input.bed> <output.lifted.bed> <output.unlifted.bed>
sh cLifted_interval.sh <chain_file> <input.bed> <output.lifted.bed> <output.unlifted.bed>
sh lLifted_interval.sh <chain_file> <input.bed> <output.lifted.bed> <output.unlifted.bed>
The command can also be used for other interval data, such as DNA targeted sequencing panel.
sh cLifted_interval.sh <chain_file> <input.bed> <output.lifted.bed> <output.unlifted.bed>
sh lLifted_interval.sh <chain_file> <input.bed> <output.lifted.bed> <output.unlifted.bed>
sh Lifted/bin/lLifted_CpG.sh hg19.fullCpGs.bedGraph hg19ToHg38.over.chain lo19to38.fullCpGs.lLifted.bedGraph e.l.unmap
sh Lifted/bin/cLifted_CpG.sh hg19.fullCpGs.bedGraph hg19ToHg38.over.chain lo19to38.fullCpGs.cLifted.bedGraph e.c.unmap
liftOver hg19.fullCpGs.bedGraph link/to/file/hg19ToHg38.over.chain lo19to38.FullCpGs.ucsc.bedGraph unlo19to38.bedGraph
segment_liftover -l link/to/liftOver -i input/folder/ -o output/folder/ -c link/to/hg19ToHg38.over.chain -si hg19.fullCpGs.bedGraph -so lo19to38.fullCpGs.segment.bedGraph
Figure 5. Comparison of Lifted (cLifted and lLifted) and UCSC liftOver, segment_liftover on Full CpGs (WGBS)
sh Lifted/bin/lLifted_interval.sh ESR1.hg19ID.bedGraph hg19ToHg38.over.chain ESR1.lo19to38.lLifted.bedGraph e.l.unmap
sh Lifted/bin/cLifted_interval.sh ESR1.hg19ID.bedGraph hg19ToHg38.over.chain ESR1.lo19to38.cLifted.bedGraph e.c.unmap
liftOver ESR1.hg19ID.bedGraph link/to/file/hg19ToHg38.over.chain ESR1.lo19to38.ucsc.bedGraph unlo19to38.ESR1.bedGraph
segment_liftover -l link/to/liftOver -i input/folder/ -o output/folder/ -c link/to/hg19ToHg38.over.chain -si ESR1.hg19ID.bedGraph -so ESR1.lo19to38.segment.bedGraph
Figure 6. Comparison of Lifted (cLifted and lLifted), UCSC liftOver and segment_liftover on a ESR1 ChIP-seq sample
