Overview¶
Gene fusion is one of the most common somatic alterations that plays an important role in tumorgenesis. Well-known examples include the intra-chromosomal TMPRSS2-ERG fusions in prostate cancer, and the inter-chromosomal BCR-ABL fusions in chronic myelogenous leukemia (CML). With the advent of next generation sequencing technologies especially RNA-seq and the development of dozens of fusion detection tools, most recurrent gene fusions in common cancers have been identified. These fusion are cataloged in databases such as COSMIC , FusionGDB , FusionHub, ChimerDB and TumorFusions ).
To facilitate molecular testing, we developed FusionVet (Fusion Visualization and Evaluation Tool) to quickly (and accurately) examine if a gene fusion with clinical significance exists in a particular sample or not.
Installation¶
You will need pip to install FusionVet. Pip is already installed if your Python3 version >= 3.4. Otherwise, follow this instruction to install pip. Use this command to install FusionVet and its dependency packages.
$ pip3 install git+https://github.com/liguowang/fusionvet.git
Alternatively, it is also available on PyPI.
$ pip3 install fusionvet
Usage¶
- Options:
--version show program’s version number and exit -h, --help show this help message and exit -b INPUT_BAM, --bam=INPUT_BAM Input BAM file. The BAM file should be sorted and indexed using SAMtools (http://samtools.sourceforge.net/). (mandatory) -c INPUT_CHIMERAS, --chimeras=INPUT_CHIMERAS Fusion file. This file can be 6 columns (chr1 start1 end1 chr2 start2 end2) or 8 columns (chr1 start1 end1 name1 chr2 start2 end2 name2) separated by Tab or Space. Lines starting with ‘#’ will be ignored. (mandatory) -o OUTPUT_FILE, --output=OUTPUT_FILE Prefix of output files. Four files will be created including “prefix.fusion.sorted.bam”, “prefix.fusion.bed”, “prefix.fusion.interact.bed” and “prefix.fusion.summary.txt”. (mandatory) -q MAP_QUAL, --mapq=MAP_QUAL Mapping quality cutoff. default=30 -t, --track-header If set, add “track line” to the BED file. -k, --keep-unknown-mapq If set, keep alignments with unknown mapping quality (i.e., MAPQ = 255).
Input files¶
FusionVet needs two types of input files.
gene fusion file¶
The gene fusion file is a plain text file with 8 columns separated by space or tab (The first 4 columns describe the “chrom”, “transcription_start”, “transcription_end” and “symbol” of gene-1, the other 4 columns describe the same information for gene-2. Below example file defines two fusions:
chr21 39739182 40033704 ERG chr21 42836477 42880085 TMPRSS2
chr14 38033152 38033701 EST14 chr7 13930855 14031050 ETV1
Output files¶
FusionVet generates 5 files
- prefix.fusion.sorted.bam
- prefix.fusion.sorted.bam.bai
- prefix.fusion.bed
- prefix.fusion.interact.bed
- prefix.fusion.summary.txt
prefix.fusion.sorted.bam¶
BAM file containing the chimeric reads supporting gene fusions. Comparing to the orignal BAM file, two additional tags are added to each alignment record: FN (Fusion Name) and SR (Supporting Read)
- SR:i:1 : Fusion was supported by split read
- SR:i:2 : Fusion was supported by paired reads
- SR:I:3 : Fusion was supported by both split read and paired reads.
$ samtools view out.fusion.sorted.bam | head -10
UNC13-SN749:172:D101FACXX:8:1104:12580:173001/1 99 chr21 39775575 66 48M = 42879910 -3104288 CTTTCACCGCCCACTCCAGCCACTGCCGCACATGGTCTGTACTCCATA CCCFFFFFHHHHHJJJJIIJJIJJJJIJIIJJJJJFHGJGFHIHIJJJ RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:1104:4678:34964/2 163 chr21 39817326 66 48M = 42879890 -3062517 CCTTGAGCCATTCACCTGGCTAGGGTTACATTCCATTTTGATGGTGAC CCCFFFDFHHHHBGHIJJJJJJIIJ?GIIGIJGGGIJJJJJJJJIFDG RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:1208:10044:4367/1 99 chr21 39817340 66 48M = 42880015 -3062628 CCTGGCTAGGGTTACATTCCATTTTGATGGTGACCCTGGCTGGGGGTT CCCFFFFFHHHFFIJJIJJJIIJJJJIIJJHJJJJIJJJIJJJJIJI> RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:1301:12176:174226/2 163 chr21 39817361 66 48M = 42879922 -3062514 TTTTGATGGTGACCCTGGCTGGGGGTTGAGACAGCCAATCCTGCTGAG BCCFFFFFHFHHHJJJJJJJJJJJJFHIIIJJIIJJJJJJJJIJIJJJ RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:2201:10011:20671/1 99 chr21 39817379 66 48M = 42879951 -3062525 CTGGGGGTTGAGACAGCCAATCCTGCTGAGGGACGCGTGGGCTCATCT CCCFFFFDHHHGHJJJJJJJJJJJIIJJJJJJIJIJGHHHFFFDEEEE RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:1108:17583:42031/2 163 chr21 39817384 65 48M = 42880007 -3062576 GGTTGAGACAGCCAATCCTGCTGAGGGACGCGTGGGCTCATCTTGGAA ?@;BDFDABFFDHHAFHHGHIIIJGIIJGIAE?@6;FGH@DDCC@CA# RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:2302:8715:52295/1 99 chr21 39817385 66 48M = 42879932 -3062500 GTTGAGACAGCCAATCCTGCTGAGGGACGCGTGGGCTCATCTTGGAAG CCCFFFFFHHHHHJJJJJHJJJJJJJJJJJJFHIJIIJGIJIJIIJIJ RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:2305:11177:45091/1 99 chr21 39817385 66 48M = 42880014 -3062582 GTTGAGACAGCCAATCCTGCTGAGGGACGCGTGGGCTCATCTTGGAAG B@CFFFFFHHHHHJJJJJJJJJJJJJIJJJJHJJJJJJJJJJJJIJJG RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:2306:12796:14838/2 163 chr21 39817391 53 48M = 42879889 -3062451 ACAGCCAATCCTGCTGAGGGACGCGTGGGCTCATCTTGGAAGTCTGTA @CCFFFFFHHHGHJJJJJJJJJJJJHGIJIJJJJJJJJIIIJHHJ### RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:1 SR:i:2 FN:Z:ERG--TMPRSS2
UNC13-SN749:172:D101FACXX:8:1308:12672:71749/1 99 chr21 39817394 66 48M = 42880007 -3062566 GCCAATCCTGCTGAGGGACGCGTGGGCTCATCTTGGAAGTCTGTCCAT ?@@FDDDFADF?D@AAB?ACGAHHEHG@BFHIGHBB=8=88@C=@@CE RG:Z:120508_UNC13-SN749_0172_AD101FACXX_8_CGATGT IH:i:1 HI:i:1 NM:i:0 SR:i:2 FN:Z:ERG--TMPRSS2
prefix.fusion.sorted.bam.bai¶
The index file of prefix.fusion.sorted.bam
prefix.fusion.bed¶
This is standard BED12 format file. Paired reads are merged into a single BED entry. This file can be uploaded to UCSC genome browser to visualize intra-chromosomal fusions. This is useful to identify the fusion point. If this file is too large to upload to UCSC genome browser directly, you could try to convert this BED file into bigBed file (using the bedToBigBed program) following this instruction.
prefix.fusion.interact.bed¶
This is Interact format file. This file can be uploaded to UCSC genome browser to visualize both intra-chromosomal and inter-chromosomal fusions. If this file is too large to upload to UCSC genome browser directly, you could try to convert this Interact file into bigInteract file (using the bedToBigBed program) following this instruction.
Intra-chromosomal fusions will be visualized as below (Note the two breaking points on ERG gene). Toggle between full display mode and pack/squish display mode help identify the exact breaking point(s).
Inter-chromosomal fusions will be visualized as below. Toggle between full display mode and pack/squish display mode help identify the exact breaking point(s).
prefix.fusion.summary.txt¶
Report the total number of supporting RNA fragments (split reads + read pairs) for each fusion.
Sample_ID ERG--TMPRSS2
Tumor_RNA_TCGA-HC-7819-01A-11R-2118-07.bam 48
Performance¶
Speed¶
FusionVet is very efficient. It took about 1 second to examine 1 fusion in a typical TCGA BAM file (7.1 Gb, 184 million reads)
Comparison to other tools¶
We used FusionVet to detect ERG-TMPRSS2 fusion from the 333 TCGA prostate cancer samples. A sample is called ERG-TMPRSS2 fusion positive if it has two or more supporting fragments.
We then compare FusionVet result to:
- FusionSeq-HighSens (Sboner et al., Genome Biology, 2010)
- MapSplice (Wang et al., Nucleic Acids Res. 2010)
- DEEPEST (Dehghannasiri et al., PNAS 2019)
- PRADA (tumorfusions.org) (Hu et al., Nucleic Acids Res. 2018)
We chose FusionSeq-HighSens and MapSplice because they were used in the original TCGA Cell paper. We chose DEEPEST and PRADA because they were newly developed and have demonstrated superior performance to other tools.
It has been found that ERG expression is significantly increased in fusion positive samples through the TMPRSS2 (an androgen responsive gene) mediated over expression (Tomlins et al., Science, 2005). Therefore, we used ERG expression as an indirect measurement of the authenticity of ERG-TMPRSS2 fusions.
FusionVet vs FusionSeq/MapSplice
FusionVet vs DEEPEST
FusionVet vs PRADA(TumorFusions.org)
