https://github.com/STAR-Fusion/STAR-Fusion
Raw File
Tip revision: e522c236a1b068ab1217855209b4706e2f8ef153 authored by Brian Haas on 27 October 2023, 18:40:42 UTC
fusionfilter update
Tip revision: e522c23
ChangeLog
STAR-Fusion Release v1.13.0, Oct 27, 2023
	- uses STAR aligner 2.7.11a
	- includes updated FusionInspector 2.9.0
	- chimeric read alignments are included in the main bam file.
	- submodule updates
	
STAR-Fusion Release v1.12.0, Jan 18, 2023
	- reverting FFPM calculation to using total reads (not unique reads as v1.11.1-only) as denominator.
	- use STAR one-pass alignment by default as faster and provides fusion results based not influenced by other reads in the sample.

	
#STAR-Fusion Release v1.11.1, Oct 3, 2022  ** retracted, reverting to total reads based FFPM calc for now **
#	- incorporates FusionInspector 2.8.0
#	- by default, FFPM calculations performed based on uniquely aligned reads.
#	    If --no_remove_dups, FFPM calculations performed using total aligned reads (including multi-mapped).
#	    For FFPM calculations based on total reads in the sample (original behavior from release 1.11.0 and earlier),
#	    set the flag parameter --FFPM_via_total_reads

	
STAR-Fusion Release v1.11.0, June 13, 2022
	- incorporates FusionInspector 2.7.0
	- provides option for STAR one-pass mode, for avoiding input-dataset-specific effects on results.
	- updates for improved reproducibility
	- added multithreading for chimeric read mapping to genes for faster execution times.



STAR-Fusion Release v1.10.1, July 23, 2021
	-incorporates FusionInspector v2.6.0
	-added option --misc_FI_opts to allow passing additional FI-specific options on to FusionInspector
	-Docker updated to use FusionInspector as the base image.

	
STAR-Fusion Release v1.10.0, Mar 6, 2021
	-includes FusionInspector v2.5.0
	-incorporates STAR 2.7.8a and updated ctat genome libs.
		-FusionAnnotator includes DEEPEST-fusion catalog
	


STAR-Fusion Release v1.9.1, Aug 10, 2020
	-includes FusionInspector v2.3.1
	-improvements for single cell data processing
	-sync max mate distance info w/ FusionInspector
	-added --max_sensitivity  and --full_Monty settings
	



STAR-Fusion Release v1.9.0, Apr 8, 2020
	-incorporates EM for assigning spanning reads to alt splice fusion isoforms
	-by default, yields single fusion call per breakpoint (matters when genes overlap)
	-incorporates FusionInspector v2.3.0 w/ pfam domain visualizations.
	-ctat genome lib improvements for IGH fusion calls


STAR-Fusion Release v1.8.1, Oct 18, 2019
	-upgraded FusionInspector to v2.2.1, reduced RAM usage


STAR-Fusion Release v1.8.0, Oct 1, 2019
	-requires STAR aligner 2.7.2b
	-bugfix for RT-artifact removal, no longer case sensitive on splice dinucleotides
	-allow exceptions to requirement of breakpoint-defining reads such as IGH--CRLF2 where breakpoints may be ill defined given complex structural rearrangements.
	-ctat genome lib updates:
		-use Dfam-based masking of cds-plus database prior to blast & homology detection. (this is similar to the earlier repeatmasker approach used in earlier versions of STAR-Fusion.  Because of repbase licensing requirements, its more straightforward to just use Dfam with nhmmer for identifying human-homologous repeat elements and softmask them prior to blast.  This was found to be particularly relevant to excluding blast-homology detection between BRAF and AKAP9 in gencode v31 where additional transcript annotations in both genes extend into Alu elements and would otherwise be detected as false homology.  The BRAF/AKAP9 fusion was falsely excluded by the previous STAR-F release, and now properly recaptured again in this release).
		-human-specific refinements: IGH and IGL super-loci now include intergenic regions; included upstream regions as targets for CRLF2, DUX4, and MALT1 known to fuse w/ IGH in cancer; hard-masking DUX4-paralogs and pseudogenes that confound IGH--DUX4 fusion calling (ideas inspired by FusionCatcher so certain important known cancer-relevant fusions aren't missed, even when the data likely stem from contaminating DNA rather than true RNA-based reads).


STAR-Fusion Release v1.7.0, August 17, 2019
	-requires newly released STAR aligner v2.7.2a, where we updated the chimeric junction reporting, requiring chimeric read alignments to have higher alignment scores than the non-chimeric read alignment counterparts. The chimeric junction output file includes all the alignment scores and new column headers to facilitate further updates.
	-now filtering likely reverse transcription artifacts - defined as those lacking canonical splicing dinucleotides at transcript chimeric breakpoints, not involving known cancer fusions, and not involving targets known to involve non-canonical breakpoints.
	-now only reporting a single fusion pair report per breakpoint. In the case there are multiple annotations overlapping the breakpoint, we report the one with the highest evidence support.
	-lowered Star-F memory requirements by streaming the chimJ file rather than putting it all into memory.
	-disabled simple homolog filtering during initial read alignment filtering steps - postponing filtering to later fusion filtering stage.

STAR-Fusion Release v1.6.0, April 2, 2019
	-requires min of STAR version 2.7.0f
	-Annot filter retains cosmic and mitelman fusions even if found in the red herrings list
	-updated CTAT genome libs and fusion annot lib, incorporates latest gtex, tcga, ccle results
	-includes latest FusionInspector w/ igv-reports stand-alone dynamic html views
	
	

	
STAR-Fusion Release v1.5.0
	-removed BBEmerge
	-require min of STAR version 2.6.1
	-leverages updated Chimeric junction out files with embedded metadata
	

STAR-Fusion Release v1.4.0
	-works with multimapping fusion evidence. Default: --STAR_chimMultimapNmax 10
	-automatically integrates overlapping mate merging for short frags or long rna-seq reads. (--bbmerge no longer required for FFPE, but still retaining the --bbmerge option for now).
	-requires at least STAR version STAR_2.6
	-minor bug fixes

	
	STAR-Fusion Release v1.3.1

	This is a major update and includes the following changes:

	-overall structure of intermediate outputs makes it more straightforward to follow the fusion evidence and fusion prediction filtering and troubleshooting (to be well described in updated documentation soon).

	-the human fusion annotation database was expanded to include some new resources, all described here:
	https://github.com/FusionAnnotator/CTAT_HumanFusionLib/wiki

	-the 'red herring' category of fusions (see above wiki) are now by default auto-filtered from the final output. You can turn off annotation-based filtering with: STAR-Fusion --no_annotation_filter

	-fusions are automatically filtered out based on minimum fusion evidence requirements set based on the number of input reads. A minimum of 0.1 fusion fragments per million total fragments (FFPM) are required as a default (set with: STAR-Fusion --min_FFPM)


STAR-Fusion Release v1.2.0
	Added --bbmerge : runs STAR w/ paired-end reads but then also runs STAR using 'super-reads' assembled via the bbmerge software.  This helps in cases where fragments are short and/or the paired-end reads overlap.
	Bugfixes related to inframe vs. frameshift detection wrt setting coding phases of exons.
	Removed filtering of blast-pairs where sequences are shared due to physical overlap on the genome (false negatives).
	FusionInspector 'inspect' and 'validate' modes write to corresponding separately named directories.
	Plug-n-play ctat genome libs are updated wrt removing blast pairs for overlapping genes and CDS phasing information for reference proteins. (GRCh37_gencode_v19_CTAT_lib_Nov012017 and GRCh38_gencode_v26_CTAT_lib_Nov012017)
	Including new CTAT genome lib for Mouse v. M15 to support mouse cancer studies.

	Instead of downloading the new ctat genome lib, users can patch the earlier: GRCh37_gencode_v19_CTAT_lib_July192017 or GRCh38_gencode_v26_CTAT_lib_July192017 involving 3 steps:
	   cd /path/to/existing/ctat_genome_lib_dir
           STAR-Fusion/FusionFilter/util/index_blast_pairs.remove_overlapping_genes.pl .
	   STAR-Fusion/FusionFilter/util/build_prot_info_db.pl --gtf ref_annot.gtf --genome_fa ref_genome.fa --out_prefix ref_annot

	
STAR-Fusion Release v1.1.0
	Much new functionality is included in this release v1.1.0:
	Note: v1.1.0 requires a new CTAT_genome_lib: https://data.broadinstitute.org/Trinity/CTAT_RESOURCE_LIB/
	annotating fusions based on known cancer fusions (--annotate)
	exploring the effects on fused coding regions, reconstructing candidate fusion transcripts and fusion proteins (--examine_coding_effect)
	extracting the fastq reads that provide the evidence for fusion transcripts (use --extract_fusion_reads)
	perform 'in silico validation' of fusion transcripts using FusionInspector ( --FusionInspector <inspect|validate> )
	de novo reconstruct fusion transcripts using Trinity (--denovo_reconstruct)


Version 1.0.0 (Feb 1, 2017)
	including dinucleotide splice sites in the report
	incorporate FFPM values whenever fastq reads are given, regardless of using the chimeric junction files direction.
	uses samtools to extract sequence regions from the genome, rather than putting the entire genome into memory.
	

Version 0.8.0 (July 28, 2016):
	FFPM stats computed correctly when using gzipped fastq files
	upgraded FusionFilter: better takes into account overlapping genes and no longer (inadvertently) filters them out as candidate paralogs.
	
Version 0.7.0 (Mar 5, 2016)
	Update to FusionFilter v0.1.0
	Require long double anchor support for those fusions only supported by breakpoint/split reads and no spanning fragments, reduces FP rate.

Version 0.6.0 (Dec 16, 2015)
	including FFPM calculations in output summary.

Version 0.5.4 (Nov, 10, 2015)
	update to FusionFilter, no 2-pass in genome-generate function

Version 0.5.3 (Nov, 6, 2015)
	removed duplicate PE alignments supporting fusions, decreases FP rate dramatically on samples w/ high PCR dups, w/ little effect on other 'good' samples w/ low DUP rate.

Version 0.5.2 (Oct 26, 2016)
	Bugfix for paired cigar-string alignment parsing leading to proper grouping of junction counts.

Version 0.5.1 (Oct 10, 2015)
     added back capability to run STAR-Fusion from an existing Chimeric.out.junctions file from an earlier STAR run.
     Use parameter:  --chimeric_junction|J or (--left_fq, --right_fq)
back to top