Fcirc is a pipeline for exploring linear transcripts and circRNAs of known fusions based on RNA-Seq data. Known fusion genes are from the multiple databases (COSMIC, ChimerDB, TicDB, FARE-CAFE and FusionCancer) or user-added gene-pairs. It costs less time to find fusions with higher sensitivity than existing methods for detecting fusions. The steps of Fcirc are as follows:

Fcirc is written in python3, requiring HISAT2 for aligning reads, samtools for selecting reads and python packages numpy, scipy, pysam.

For running Fcirc a computer with the following configuration is needed:

• minimum 8 GB of RAM (aligning to human genome reference)
• 1 CPU (minimum)

    git clone https://github.com/WangHYLab/fcirc

• hisat2 http://ccb.jhu.edu/software/hisat2/index.shtml
• samtools http://www.htslib.org/download/
• some python packages that can be installed by pip are as follows:

pip install -r requirements.txt

or

pip install numpy
pip install scipy
pip install pysam
pip install cutadapt

Make sure that hisat2 and samtools are added to environment variables so that Fcirc can invoke them.

• The genome resource is hisat2 index, which can be downloaded from hisat2 websites[http://ccb.jhu.edu/software/hisat2/index.shtml]. For human fusion transcript detection, it's recommended to use genome_tran of GRCh38 or GRCh37. It can also be finished with FASTA sequence file and annotation GTF file by hisat2 script.

• Known fusion-pairs can be downloaded from Github page and bipartite fusions index can be built by hisat2-build in reference_fusion_info directory as follows:

cd reference_fusion_info
hisat2-build fusiongenes_ref_U.fa fusiongenes_ref_U
hisat2-build fusiongenes_ref_V.fa fusiongenes_ref_V

• (optional) Add gene pairs of fusions you are interested in.

python3 build_graph.py --genome absolute__to_genome --gtf absolute__to_gtf --tab absolute_path_to_fusionpairs_table
    

# Warning: gene pairs joined with'--' not '-' should be placed in the first column of fusionpairs_table

The input data shall be single-end or paired-end RNA-Seq in FASTQ format, which can be raw data or trimmed data.

Fcirc can be run with a simple command line.

python fcirc.py [options] -x <ht2-trans-idx> -f <ht2-fusion-idx-dir> -c <fusion-genes-coordinates> {-1 <fastq1> | -1 <fastq1> -2 <fastq2>} 

Arguments can be used as following:

    Required:
        -x <ht2-trans-idx>, --trans_idx <ht2-trans-idx>
            transcription index filename prefix (minus trailing .X.ht2)
        -f <ht2-fusion-idx-dir>, --fusion_idx_dir <ht2-fusion-idx-dir>
            fusion index directory (contains fusiongenes_ref_U and fusiongenes_ref_V)
        -c <fusion-genes-coordinates> --fusion_genes_coord
            fusion genes coordinates file (defalut: fusion_genes_coordinate.txt)    
        -1 <fastq1>, --file1 <fastq1>
            fastq file 1 (single-end pattern: only -1)
        -2 <fastq2>, --file2 <fastq2>
            fastq file 2 (paired-end pattern: -1 and -2, files should be like -1 xxx_1.fastq -2 xxx_2.fastq)

    Optional:
        -q <quality_val>
           the minimum phred qulaity of read(default:0)
        -o <output_dir>, --output <outout_dir>
            output file directory (default: .)
        -t <int>, --thread <int>
            number of hisat2 alignment and pysam filter threads to launch (default: 1)    

    Others:
        -h, --help
            help information  
        -v, --version
            version information 

The output includes:

• fusion information
• f-circRNA information

1. Fusion information is stored in a file 'fusion_results.tsv' as the following format:

#Fusion_Name	5'Gene	3'Gene	5'Gene_chr	5'Gene_strand	3'Gene_chr	3'Gene_strand	5'Gene_BreakPoint_Pos	3'Gene_BreakPoint_Pos	5'and3'_Common_Breakpoint_Seq	BreakpointReads_Count	BreakpointReads	BreakpointStrand_Count(+,-)	ScanningReads_Count	ScanningReads	ScanningStrand_Count(+,-)	P-Value
PML-RARA	PML	RARA	15	+	17	+	74023408	40348313	.	117	SRR3239817.48728782,SRR3239817.46047306,SRR3239817.46553524,SRR3239817.16929141,SRR3239817.19547854,SRR3239817.24567755,SRR3239817,......

The description of each column:
#Fusion Name - - The name of the fusion
5'Gene - - The gene encoding the 5' end of the fusion transcript
3'Gene - - The gene encoding the 3' end of the fusion transcript
5'Gene_chr- - The chromosome of 5'end gene
5'Gene_strand- - The strand of 5'end gene
3'Gene_chr -- The chromosome of 3'end gene
3'Gene_strand- - The strand of 3'end gene
5'Gene BreakPoint Pos - - The position of the breakpoint for the 5' end of the fusion transcript
3'Gene BreakPoint Pos - - The position of the breakpoint for the 3' end of the fusion transcript
5'and 3'Common Breakpoint Seq - - The same sequence at the breakpoint of the 3' end of the gene and the 5' end of the gene
BreakpointReads Count - -The number of reads spanning the fusion breakpoint
BreakpointReads - -The reads spanning the fusion breakpoint
BreakpointStrand Count(+,-) - - The number of reads located in forward strand and reverse strand respectively
ScanningReads Count(+,-) - - The number of pair of reads are located on both sides of the breakpoint
ScanningReads- - The reads located on both sides of the breakpoint
ScanningStrand_Count(+,-) -- The number of Scanning reads located in forward strand and reverse strand respectively
P-Value - - A p value indicating if reads around the breakpoint are evenly distributed

2. FcircRNA information is stored in a file 'fcircRNA_results.tsv' as the following format:

#FcircRNA_NO	Fusion Name	Backsplice_start	Backsplice_end	Fusion5'_BreakPoint_Pos	Fusion3'_BreakPoint_Pos	Support_FcircRNA_Reads_Count	FcircRNA_Strand_Count(+, -)	Support_FcircRNA_Reads
No_1	PML-RARA	15:74023268:+	17:40351924:+	15:74023408:+	17:40348313:+	1	0,1	SRR3239817.23906640
No_2	PML-RARA	15:73998438:+	17:40352058:+	15:74023408:+	17:40348313:+	3	0,3	SRR3239817.6429653,SRR3239817.3386413,SRR3239817.31829112
No_3	PML-RARA	15:73998328:+	17:40354455:+	15:74023408:+	17:40348313:+	1	0,1	SRR3239817.3123010
No_4	PML-RARA	15:73998193:+	17:40352044:+	15:74023408:+	17:40348313:+	5	0,5	SRR3239817.29876711,SRR3239817.36732283,SRR3239817.47058005,SRR3239817.32495621,SRR3239817.13611951
No_5	PML-RARA	15:73998454:+	17:40352406:+	15:74023408:+	17:40348313:+	1	0,1	SRR3239817.28808693
No_6	PML-RARA	15:74022909:+	17:40355327:+	15:74023408:+	17:40348313:+	3	0,3	SRR3239817.42010789,SRR3239817.11495312,SRR3239817.33451057
......
......
...

The description of each column:
#FcircRNA_NO - - The id of fusion circRNA
Fusion Name - - The name of fusion gene
Backsplice start - - The starting position of back-spliced end
Backsplice end - - The end position of back-spliced end
Fusion5'_BreakPoint_Pos - - The position of fusion breakpoint on 5'end
Fusion3'_BreakPoint_Pos - - The position of fusion breakpoint on 3'end
Support_FcircRNA_Reads_Count - - The number of reads supporting the f-circRNA
FcircRNA_Strand_Count(+, -) - - The number of reads supporting f-circRNA on positive and negative strand
Support FcircRNA Reads - - The reads supporting the f-circRNA

You can start this pipeline using a testing RNA-Seq data, whose reads are partially from a RNA-Seq dataset SRR3239817 (NCBI SRA database), for an acute leukaemia cell line NB4.

python fcirc.py -t 4 -o fcirc_out -x transcriptome_HISAT2_index_path -f known_fusion_directory_path -1  test_fastq_path

It costs few minutes. If it runs successfully, some log information will be printed as following:

[2020-04-29 11:00:32] Finish mapping reads to transcription!
[2020-04-29 11:00:33] Finish mapping reads to fusion references U!
[2020-04-29 11:00:33] Finish mapping reads to fusion references V!
[2020-04-29 11:00:33] Finish dropping unmapped read in fusion references U and V!
Find 215 Reads in U! 215 Reads in V!
[2020-04-29 11:00:34] Finish filtering fusion-related reads in fusion references U and V!
[2020-04-29 11:00:36] Finish mapping reads to inferred fusion references!
Find 22 kind(s) of fcircRNAs!
[2020-04-29 11:00:36] Finish all!See the result in 'fcircRNA_results.tsv'!