Usage¶

Basic usage¶

usage: CIRI-long [-h] [-v] {call,collapse} ...

positional arguments:
  {call,collapse}  commands

optional arguments:
  -h, --help       show this help message and exit
  -v, --version    show program's version number and exit

CIRI-long have two main functions, including (1) candidate circRNAs identification and (2) isoform collapsing.

Step1. circRNA identification¶

Basic options¶

usage: CIRI-long call [-h] [-i READS] [-o DIR] [-r REF] [-p PREFIX] [-a GTF] [--canonical] [-t INT] [--debug]

optional arguments:
  -h, --help            show this help message and exit
  -i READS, --in READS  Input reads.fq.gz
  -o DIR, --out DIR     Output directory, default: ./
  -r REF, --ref REF     Reference genome FASTA file
  -p PREFIX, --prefix PREFIX
                        Output sample prefix, (default: CIRI-long)
  -a GTF, --anno GTF    Genome reference gtf, (optional)
  -c CIRC, --circ CIRC  Additional circRNA annotation in bed/gtf format,
                        (optional)
  -t INT, --threads INT
                        Number of threads, (default: use all cores)
  --debug               Run in debugging mode, (default: False)

NOTE:

A bwa index for reference genome is required, please use bwa index command to generate bwa index before running CIRI-long.

Example Usage:¶

Demo dataset can be downloaded from the GitHub release

# Download demo dataset
wget https://github.com/bioinfo-biols/CIRI-long/releases/download/v0.6-alpha/CIRI-long_test_data.tar.gz

# Decompress demo dataset
tar zxvf CIRI-long_test_data.tar.gz
cd test_data

# Build bwa index before running CIRI-long
bwa index -a bwtsw mm10_chr12.fa mm10_chr12.fa

# Run CIRI-long to identify circular reads from sequencing reads
CIRI-long call -i test_reads.fa \
               -o ./test_call \
               -r mm10_chr12.fa \
               -p test \
               -a mm10_chr12.gtf \
               -t 8

Output Files¶

The output directory should have the following structure:

test_call
├── test.cand_circ.fa
├── test.json
├── test.log
├── test.low_confidence.fa
└── tmp
    ├── ss.idx
    ├── test.ccs.fa
    └── test.raw.fa

1 directory, 7 files

Using non-canonical splice signals¶

If you would like to use other splice signals, please modify the dict SPLICE_SIGNAL in align.py in format: {(5’SS, 3’SS): Priority}

Default configuration:

SPLICE_SIGNAL = {
    ('GT', 'AG'): 0,  # U2-type
    ('GC', 'AG'): 1,  # U2-type
    ('AT', 'AC'): 2,  # U12-type
    ('GT', 'AC'): 2,  # U12-type
    ('AT', 'AG'): 2,  # U12-type
}

Using additional circRNA annotations¶

From version v1.0.2, CIRI-long call also provide additional circRNA annotations in BED/GTF format for BSJ correction with --circ option. CircRNA annotations can be downloaded from circAtlas or other databases. The GTF-format output of CIRIquant is also supported.

NOTE: If using results from other tools/databases, please make sure the coordinate system is compatible with our CIRI-series tools:

The coordinate system of circRNAs is different in most circRNA tools. For instance, if a circRNA is derived from chr1:1000-2000, it should be reported as chr1:1000-2000 in CIRI-series and some tools (DCC/KNIFE/Mapsplice), but reported as chr1:999-2000 in other tools (CIRCexplorer2/UROBORUS/circRNA_finder/find_circ).

Thus, if you want to use circRNAs identified from tools in the latter group, you need to add 1 extra base to the start coordinate of circRNAs (the position with smaller coordinate regardless of the strand information), then use the altered coordinates as input.

Step2. isoform collapse¶

Basic Options¶

usage: CIRI-long collapse [-h] [-i LIST] [-o DIR] [-p PREFIX] [-r REF] [-a GTF] [--canonical] [-t INT] [--debug]

optional arguments:
  -h, --help            show this help message and exit
  -i LIST, --in LIST    Input list of CIRI-long results
  -o DIR, --out DIR     Output directory, default: ./
  -p PREFIX, --prefix PREFIX
                        Output sample prefix, (default: CIRI-long)
  -r REF, --ref REF     Reference genome FASTA file
  -a GTF, --anno GTF    Genome reference gtf, (optional)
  -c CIRC, --circ CIRC  Additional circRNA annotation in bed/gtf format,
                        (optional)
  -t INT, --threads INT
                        Number of threads, (default: use all cores)
  --debug               Run in debugging mode, (default: False)

One should provide a text file listing sample name and path to CIRI-long output files *.cand_circ.fa, seperated by space.

sample1_name /path/to/sample1/cand_circ.fa
sample2_name /path/to/sample2/cand_circ.fa

Example Usage¶

For exmaple, you can create a file name test.lst with the following content:

test ./test_call/test.cand_circ.fa

Then run CIRI-long collapse to aggregate results from one or multiple samples.

 CIRI-long collapse -i ./test.lst \
                    -o ./test_collpase \
                    -p test \
                    -r ./mm10_chr12.fa \
                    -a ./mm10_chr12.gtf \
                    -t 8

Output Files¶

The output directory should have the following structure:

test_collpase
├── test_collpase.expression
├── test_collpase.isoforms
├── test_collpase.info
├── test_collpase.log
├── test_collpase.reads
└── tmp
    ├── ss.idx
    └── test_collpase.corrected.pkl

1 directory, 6 files

Output Format¶

The main output

The main output of CIRI-long is a GTF file (e.g. test_collpase.info), that contains detailed information of circRNAs and annotation of circRNA back-spliced regions in the attribute columns

Description of each columns’s value

column	name	description
1	chrom	chromosome / contig name
2	source	CIRI-long
3	type	circRNA
4	start	5' back-spliced junction site
5	end	3' back-spliced junction site
6	score	Number of total supported reads
7	strand	strand information
8	.	.
9	attributes	attributes seperated by semicolon

The attributes containing several pre-defined keys and values:

key	description
circ_id	name of circRNA
splice_site	splicing signal of candidate circRNAs and numbers indicating shifted bases of aligned and annotated splice site. (e.g. AG-GT\|0-5)
equivalent_seq	equivalent sequence of splice site
circ_type	circRNA types: exon/intron/intergenic
circ_len	length of the major isoform of circRNA
isoform	structure of isoforms, isoforms are seperated by "\|" and circular exons are seperated by "," (e.g. 11627815-111627914,111628190-111628302\|11627815-111628302)
gene_id	ensemble id of host gene
gene_name	HGNC symbol of host gene
gene_type	type of host gene in the annotation gtf file

Expression matrix

test_collpase.expression contains the summarized expression level of circRNAs in all samples in tsv format.
test_collpase.isoforms contains the summarized isoform usage index of assembled isoforms in all samples in tsv format.
Isoform usage index = Isoform_reads / Sum of all isoforms from the same BSJ

Step3. Output visualization¶

From version v1.1.0, CIRI-long included the ‘misc/convert_bed.py’, users can convert the GTF-formatted circRNA.info to bed format, and visualize using softwares like IGV / Jbrowse2

python3 misc/convert_bed.py collapse_out/sample.info sample_circ.bed