Gene fusions play a major role in tumorigenesis, so it is crucial that Treehouse has a pipeline for detecting them. We have built a docker container that runs STAR-Fusion and filters the output against a list of known cancer fusion genes. There is also an option to run additional filters and generate de novo assembled fusion transcripts using the FusionInspector program.
REPOSITORY: ucsctreehouse/fusion
TAG: 0.2.0
The pipeline requires paired-end fastq files, the output directory, and the genome library directory. The genelist is already baked into the docker container, but there is an option to include a different genelist. Please refer to the STAR-Fusion documentation for creating a genome library. You can also find a prebuilt genome library here: http://ceph-gw-01.pod/references/STARFusion-GRCh38gencode23.tar.gz
Wraps STAR-Fusion program and filters output using FusionInspector.
optional arguments:
-h, --help show this help message and exit
--left-fq R1 Fastq 1
--right-fq R2 Fastq 2
--output-dir OUTPUT_DIR
Output directory
--genome-lib-dir GENOME_LIB_DIR
Reference genome directory (can be tarfile)
--CPU CPU Number of jobs to run in parallel
--genelist GENELIST
--skip-filter Skips gene-list filter
-F, --run-fusion-inspector
Runs FusionInspector on STAR-Fusion output
--star-fusion-results STAR_FUSION_RESULTS
Skips STAR-Fusion and runs FusionInspector
--save-intermediates Does not delete intermediate files
--root-ownership Does not change file ownership to user
--test Runs the pipeline with dummy files
--debug Prints tool command line arguments
docker run -it --rm -v `pwd`:/data ucsctreehouse/fusion:0.1.0 \
--left-fq 1.fq.gz \
--right-fq 2.fq.gz \
--output-dir fusion_output \
--genome-lib-dir STARFusion-GRCh38gencode23
--run-fusion-inspector
STAR-Fusion Output
- STAR-Fusion candidates and fusion fragments per million mapped read values
- star-fusion-non-filtered.final
- star-fusion-non-filtered.final.bedpe
- Fusions where donor and acceptor genes passed the gene-list filter
- star-fusion-gene-list-filtered.final
- star-fusion-gene-list-filtered.final.bedpe
FusionInspector Output
- FusionInspector candidates and fusion fragments per million mapped read values where donor and acceptor passed gene-list filter
- fusion-inspector-results.final
- fusion-inspector-results.final.bedpe
- IGV input files for visualization
- FusionInspector.fa
- FusionInspector.gtf
- FusionInspector.junction_reads.bam
- FusionInspector.spanning_reads.bam
FusionInspector predictions with a joint FFPM > 0.1 are considered significant.