RNA Fusion Detection and Quantification using STAR.
Post-alignment run times are typically <20 minutes using 4 threads. Development is still ongoing and several features are currently in the works. DNA breakpoint detection is still experimental.
This package is tested under Linux using Python 2.7, 3.4, 3.5, and 3.6.
You can install from Pypi. Please use a recent version of pip and cython:
pip install -U pip pip install -U cython pip install starseqr
Or build directly from Github by cloning the project, cd into the directory and run:
python setup.py install
Or from Docker:
docker pull eagenomics/starseqr
Or from Bioconda:
conda install -c bioconda starseqr
- Additional Requirements
- biobambam2(https://github.com/gt1/biobambam2) or conda install -c bioconda biobambam
- STAR(https://github.com/alexdobin/STAR). Must use >2.5.3a. conda install -c bioconda star
- Velvet(https://github.com/dzerbino/velvet) or conda install -c bioconda velvet
- samtools(https://github.com/samtools/samtools) or conda install -c bioconda samtools
- Salmon(https://combine-lab.github.io/salmon/) or conda install -c bioconda salmon
- UCSC utils(http://hgdownload.soe.ucsc.edu/admin/exe/) or conda install -c bioconda ucsc-gtftogenepred
- gffread(http://ccb.jhu.edu/software/stringtie/dl/gffread-0.9.8c.tar.gz) or conda install -c bioconda gffread
First make sure the dependencies are installed and generate a STAR index for your reference.
RNA Index
STAR --runMode genomeGenerate --genomeFastaFiles hg19.fa --genomeDir STAR_SEQR_hg19gencodeV24lift37_S1_RNA --sjdbGTFfile gencodeV24lift37.gtf --runThreadN 18 --sjdbOverhang 150 --genomeSAsparseD 1
STAR-SEQR can perform alignment or utilize existing outputs from STAR. Note- STAR-SEQR alignment parameters have been tuned for fusion calling.
Python on OS
starseqr.py -1 RNA_1.fastq.gz -2 RNA_2.fastq.gz -m 1 -p RNA_test -t 12 -i path/STAR_INDEX -g gencode.gtf -r hg19.fa -vv
CWL
Note that --name_prefix must be a string basename in this case.
cwltool ~/path/STAR-SEQR/devtools/cwl/starseqr_v0.6.6.cwl --fq1 /path/UHRR_1_2_5m_L4_1.clipped.fastq.gz --fq2 /path/UHRR_1_2_5m_L4_2.clipped.fastq.gz --star_index_dir /path/gencodev25lift37/STAR_INDEX --name_prefix test_cwl --transcript_gtf /path/gencodev25/gencode.v25lift37.annotation.gtf --genome_fasta /path/gencodev25/GRCh37.primary_assembly.genome.fa --mode 1 --worker_threads 8
DOCKER
Note that -p must be a fully qualified path in this case.
docker run -it -v /mounts:/mounts eagenomics/starseqr:0.6.5 starseqr.py -1 /mounts/path/UHRR_1_2_5m_L4_1.clipped.fastq.gz -2 /mounts/path/UHRR_1_2_5m_L4_2.clipped.fastq.gz -p /mounts/path/test_docker -i /mounts/path/gencodev25lift37/STAR_INDEX -g /mounts/path/gencodev25/gencode.v25lift37.annotation.gtf -r /mounts/path/gencodev25/GRCh37.primary_assembly.genome.fa -m 1 -vv
A BEDPE file is produced and is compatible with SMC-RNA Dream Challenge.
Breakpoints.txt and Candidates.txt have the following columns:
| Values | Description |
| NAME | Gene Symbols for left and right fusion partners |
| NREAD_SPANS | The number of paired reads that are discordant spanning and suppor the fusion |
| NREAD_JXNLEFT | The number of paired reads that are anchored on the left side of the gene fusion |
| NREAD_JXNRIGHT | The number of paired reads that are anchored on the right side of the gene fusion |
| FUSION_CLASS | Classification of fusion based on chromosomal location, distance and strand. [GENE_INTERNAL, TRANSLOCATION, READ_THROUGH, INTERCHROM_INVERTED, INTERCHROM_INTERSTRAND] |
| SPLICE_TYPE | Classification of the fusion breakpoint. If on the exon boundary is CANONICAL, else NON-CANONICAL |
| BRKPT_LEFT | The 0-based genomic position of the fusion breakpoint for the left gene partner |
| BRKPT_RIGHT | The 0-based genomic position of the fusion breakpoint for the right gene partner |
| LEFT_SYMBOL | The left gene symbol |
| RIGHT_SYMBOL | The right gene symbol |
| ANNOT_FORMAT | The description of keys that are used in the ANNOT column. Similar to VCF FORMAT notation. |
| LEFT_ANNOT | The values described in the ANNOT_FORMAT column for the left gene breakpoint |
| RIGHT_ANNOT | The values described in the ANNOT_FORMAT column for the right gene breakpoint |
| DISTANCE | The genomic distance between breakpoints. Empty if a translocation. |
| ASSEMBLED_CONTIGS | The velvet assembly of the supporting chimeric reads |
| ASSEMBLY_CROSS_JXN | A boolean value indicating if the assembly crosses the putative breakpoint |
| PRIMERS | Primers left, right designed against the highest expressing predicted fusion transcript |
| ID | Internal notation of STAR-SEQR breakpoints. |
| SPAN_CROSSHOM_SCORE | Homology score with range of [0-1] to indicate the probability of spanning chimeric reads mapping to both gene partners |
| JXN_CROSSHOM_SCORE | Homology score with range of [0-1] to indicate the probability of junction chimeric reads mapping to both gene partners |
| OVERHANG_DIVERSITY | The number of unique fragments that fall from left anchored split-reads onto the right gene and vice-versa. |
| MINFRAG20 | The number of overhang fragments that have at least 20 bases |
| MINFRAG35 | The number of overhang fragments that have at least 35 bases |
| TPM_FUSION | Expression of the most abundant fusion transcript expressed in transcripts per million |
| TPM_LEFT | Expression of the most abundant left transcript expressed in transcripts per million |
| TPM_RIGHT | Expression of the most abundant right transcript expressed in transcripts per million |
| MAX_TRX_FUSION | Highest expressing fusion transcript. Expression corresponds to TPM_FUSION |
| DISPOSITION | Values to indicate PASS or other specific reasons for failure |
Yes! Please give us your feedback, raise issues, and let us know how the tool is working for you. Pull requests are welcome.
This project builds of the groundwork of other public contributions. Namely:
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent