INTEGRATE-Circ is a fusion junction detection tool, capable of identifying gene fusion transcripts and various isoforms that may result from these fusions, including fusion-derived circRNA (fcircRNA). INTEGRATE-Circ is an extension of the algorithm employed by INTEGRATE, which can be found here.
INTEGRATE-Circ is developed at the Christopher Maher Lab at Washington University in St. Louis.
First, clone the INTEGRATE-Circ repository:
cd PATH_TO_TOOL
git clone https://github.com/ChrisMaherLab/INTEGRATE-CIRC.git
Next, compile the tool (requires cmake version >= 2.8):
cd INTEGRATE-Circ
mkdir build
cd build
cmake ..
make
An executable file should now be located at PATH_TO_TOOL/INTEGRATE-CIRC/build/bin/Integrate-Circ. To make this file easily accessible from other locations, update the $PATH variable using export PATH="PATH_TO_TOOL/INTEGRATE-CIRC/build/bin:$PATH" at the command line.
Note that INTEGRATE-Circ cannot be compiled by some older compilers (recommend version >12 of g++). If this is not your default, you can explicitly tell cmake to use this by installing an updated version and then specifying it by changing the cmake .. command above to cmake .. -DCMAKE_CXX_COMPILER=<path_to_g++-12>.
Besides user-provided sequencing data, two input files are required for INTEGRATE-Circ.
- Reference genome fasta file
- Ensembl annotation file
Examples of each can be found in the example-data directory.
Prior to analyzing sequencing data, INTEGRATE-Circ must perform a BWT (Burrows-Wheeler transform) of the reference genome. This operation only needs to be performed once and can be done using the following command. If no output directory is specificed, output defaults to ./bwts.
Integrate-Circ mkbwt /path/to/reference.fa -dir /path/to/bwt_directory
After running the above command, sequencing analysis can be performed:
Integrate-Circ fusion \
/path/to/reference.fa \
/path/to/annotation_file.txt \
/path/to/bwt_directory \
/path/to/mapped_RNA_reads.bam \
/path/to/unmapped_RNA_reads.bam \
/path/to/mapped_tumor_DNA.bam \
/path/to/mapped_normal_DNA.bam
By default, all outputs are placed in the current working directory.
If mapped and unmapped RNA reads are in the same file, simply provide that same file twice. Both DNA files are optional. Integrate-Circ fusion -h can be run for a more complete description of parameters, which are also described below.
The INTEGRATE-circ fusion command outputs the following files:
- summary.tsv: Contains a summary of all called fusions as well as any associated splice variants, including backsplices
- exons.tsv: Contains information regarding fusion junctions, useful for creating primers for validation sequencing
- breapoints.tsv: Describes fusion breakpoints as determined by RNA and DNA data
- bk_sv.vcf: Describes fusion breakpoints in vcf format
- fusions.bedpe: Describes fusion and splice junctions in SMC-RNA bedpe format
- reads.txt: Describes information about all sequencing reads that support the identified junctions
- fcirc.txt: Descibes fcircRNAs. Columns 1 and 2 provide ID and fusion gene information. Column 3 describes the backsplice acceptor, column 4 describes the backsplice donor, column 5 descibes the 5' fusion junction and column 6 describes the 3' fusion junction. Note that in a geneA::geneB fusion, columns 3 and 5 describe locations in geneA and columns 4 and 6 describes locations in geneB.
Analysis of the example-data can be performed as follows. The example data includes simulated reads consistent with a TMPRSS2::ERG gene fusion that also creates an fcircRNA.
cd PATH_TO_TOOL
mkdir example-data/bwts
Integrate-Circ mkbwt example-data/hg19.chr21.fa -dir example-data/bwts
Integrate-Circ fusion \
example-data/hg19.chr21.fa \
example-data/hg19.chr21.annotation.txt \
example-data/bwts \
example-data/RNA.Example.AllReads.bam \
example-data/RNA.Example.AllReads.bam
By default, all outputs are placed in the current working directory. For visualization of the TMPRSS2::ERG fcircRNA, the output fcirc.txt file can be supplied as input to the most recent version of the INTEGRATE-vis tool developed by the Chris Maher Lab.
Running INTEGRATE-Circ mkbwt (without parameters) gives the following parameter options:
Integrate-Circ mkbwt (options) reference.fasta
options:
-mb integer : sequences in the reference fasta that are shorter than this value default: 10000000
are not included in the evaluation of repetitive reads.
-dir string : directory to store the BWTs. default: ./bwts
Running INTEGRATE-Circ fusion -h gives the following information:
Integrate-Circ fusion (options) reference.fasta annotation.txt directory_to_bwt accepted_hits.bam unmapped.bam (dna.tumor.bam dna.normal.bam)
options: -cfn integer : Cutoff of spanning RNA-Seq reads for fusions with non-canonical
exonic boundaries default: 3
-rt float : Normal dna / tumor dna ratio. If the ratio is less than
this value, then dna reads from the normal dna data set
supporting a fusion candidates are ignored default: 0.0
-minIntra integer : If only having RNA reads, a chimera with two adjacent
genes in order is annotated as intra_chromosomal rather than
read_through if the distance between the two genes is larger than
this value default: 400000
-minW float : Mininum weight for the encompassing rna reads on an edge default: 2.0
-mb integer : See subcommand "mkbwt"
This value can be larger than used by mkbwt. default: 10000000
-minDel int : minimum size of a deletion that can cause a fusion. default: 5000
-reads string : File to store all the reads default: reads.txt
-sum string : File to store summary default: summary.tsv
-ex string : File to store exons for fusions with canonical exonic boundaries default: exons.tsv
-bk string : File to store breakpoints default: breakpoints.tsv
-vcf string : File to store breakpoints in vcf format default: bk_sv.vcf
-fcirc string : File to store fcirc results in default: fcirc.txt
-bedpe string : File to store all fusions in SMC-RNA bedpe format default: junctions.bedpe
-bacc integer : max difference between spanning reads and annotation to decide canonical default: 1
-largeNum integer : if a gene shows greater or equal to this number, remove it from results default: 4
-sample string : sample name default: sample
-dir string : Name of directory to create for storing outputs default: INTEGRATE_Circ_output
This version of Integrate-Circ works in the following situations:
(1)having rna tumor, dna tumor, dna normal
(2)having rna tumor, dna tumor
(3)having rna tumor
Integrate-Circ will only use sequences in reference.fasta.
Chr names with and without "chr" are regarded as the same, e.g. chr1 = 1.
The rna and dna bams can be from alignments mapped to different reference files with different order of the sequences and their names with or without "chr". However, The versions should be the same, e.g. hg19. (Also, the same as in annotation.)
The tumor and normal dna bams should be mapped to the same reference file.
For rna tumor: accepted_hits.bam is a bam file containing mapped rna reads. unmapped.bam is a bam contains the not mapped rna reads. If they have been merged into one bam, just use merged.bam twice in the command line.
For dna bams: If solt-clips are provided, then Integrate-Circ is trying to search rearrangement breakpoints, otherwise, only paired reads may be included in the analysis.
If having rna normal only or having both rna and dna normal data sets. These data sets can be run to find non somatic events.
e.g. Integrate-Circ fusion -normal (options) reference.fasta annotation.txt directory_to_bwt accepted_hits.normal.bam unmapped.normal.bam (dna.normal.bam)
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent