pVACfuse logo


Fusion detection and annotation

pVACfuse accepts two types of inputs, either an annotated bedpe file with fusion information from INTEGRATE-Neo or a output directory from AGFusion (recommended).


AGFusion allows a user to annotate output files from several fusion callers using the agfusion batch command. The below example is for annotating the output from the STAR-Fusion caller but many other fusion callers are supported. For a full list see the AGFusion documentation.

agfusion batch \
-f <star_fusion_tsv> \
-a starfusion \
-db agfusion.homo_sapiens.87.db \
- <output_directory> \
--middlestar \

The --middlestar flag is required in order to use the ouput with pVACfuse. This will indicate the fusion position in the fusion peptide sequence.

The --noncanonical flag is optional and can be used to annotate the fusion with informations from all possible transcripts. By default only canonical transcripts are used.


Fusion detection will be preformed using INTEGRATE with annotations from INTEGRATE-Neo. It should be possible to start with fusions from another caller, convert the output to bedpe format, annotate the bedpe with INTEGRATE-Neo and then feed these candidates into pVACfuse.

  1. Align RNA with Tophat2 (a requirement of INTEGRATE) to obtain accepted_hits.bam and unmapped.bam

  2. (OPTIONAL) Align WGS DNA with BWA aln/sampe (NOT MEM, a requirement of INTEGRATE) to obtain tumor.dna.bam and normal.dna.bam

  3. Produce a gene annotations file with gtfToGenePred

gtfToGenePred -genePredExt -geneNameAsName2 ref.gtf ref.genePred
cut -f 1-10,12 ref.genePred > tmp.txt
echo -e "#GRCh37.ensGene.name\tGRCh37.ensGene.chrom\tGRCh37.ensGene.strand\tGRCh37.ensGene.txStart\tGRCh37.ensGene.txEnd\tGRCh37.ensGene.cdsStart\tGRCh37.ensGene.cdsEnd\tGRCh37.ensGene.exonCount\tGRCh37.ensGene.exonStarts\tGRCh37.ensGene.exonEnds\tGRCh37.ensemblToGeneName.value" > annot.txt
cat tmp.txt >> annot.txt
  1. Run INTEGRATE to obtain fusions.bedpe

Integrate fusion ref.fa annot.txt bwts accepted_hits.bam unmappeds.bam [tumor.dna.bam normal.dna.bam | tumor.dna.bam]
  1. Run INTEGRATE-Neo to obtain annotated fusions bedpe file

integrate-neo.py -t hla.optitype -f fusions.bedpe -r ref.fa -g ref.genePred -k