2.1.0

Functional changes

• No changes to pipeline output or results; this release focuses on internal refactoring and maintainability improvements.
• In rare cases, no_frame or neo_frame fusion annotations may change due to improved handling of reference annotations.

Internal & Technical Changes

Code Refactoring

• Refactored fusion annotation to modular fusionannotator.py and supporting components
• Refactored fusion parsing; fusiontoolparser.py now consumes standardized per-tool CSVs via repeated --tool and writes Detected_Fusions.csv with consistent headers
• Nextflow workflow reorganized: parsing moved to modules/04_fusionparsing.nf, annotation to modules/05_fusionannotation.nf; downstream stages renumbered
• Retrained model due to slight changes in annotations

Technical Improvements

• Robust header-based range extraction in read_selection.py for wildtype ranges
• Stricter chromosome filtering to primary contigs and strand handling in tool parsers
• Pipeline outputs aligned for ARRIBA (only fusions.tsv) and downstream consumers

Infrastructure & Dependencies

• Updated environments: conda channels switched to nodefaults
• Removed logzero dependency and related logging calls; reduced log output across scripts
• Removed legacy monolithic fusion annotation script (replaced by fusionannotator.py)
• Removed ARRIBA discarded output from pipeline (structural cleanup; discarded fusions were not used downstream before, only high-confidence calls proceed)

Development & Testing

• New utilities for annotation: gff3_to_db.py (build gffutils DB) and gtf2tsl.py (extract TSL)
• New modular fusion parsing with per-tool parsers (Arriba, STAR-Fusion, FusionCatcher, InFusion, MapSplice, SOAPfuse) producing standardized CSV via parse_tool.py
• New Nextflow parsing processes (PARSE_ARRIBA, PARSE_STAR_FUSION, PARSE_FUSION_CATCHER) and dedicated conda env (environments/fusionparsing.yml)
• Unit tests and test runner for fusion annotation module

2.0.4

Added

• Added full length protein sequence to the final output
• Specifiy computational requirements via predefined labels: single, low and medium

Changed

• Updated NextflowVersion to 24.10.1
• Updated resource management
• Fixed exon count in final output
• Fixed tool_frac column in final output
• Updated prediction model based on new results

2.0.3

Added

• Arriba v2.4.0 high confidence calls as fusion candidates
• easyquant v0.5.2 for read support requantification
• Unit/integration tests using pytest

Changed

• Fixed issue with gene names in fusion annotation script
• Updated prediction model based on new results
• Moved conversion, parsing and annotation code from the easyfuse-src package
• Removed unnecessary columns from final output

2.0.2

• Upgraded pipeline to Ensembl v110
• Updated to FusionCatcher v1.33

2.0.1

• Simplified installation and dependency management through migration of EasyFuse package to Bioconda
• Fixed bug in QC workflow

2.0.0

This release includes the following major changes:

• EasyFuse as NextFlow pipeline for increased usability, stability, and scalability
• Python code as python package outsourced to separate repository
• Internal detection tools were reduced to StarFusion and FusionCatcher
• Prediction model has been changed to EF_requant_type to not rely on specific tool features
• Overall reduced detection performance in sensitivity and precision compared to EasyFuse 1.3.7

1.3.7

• Fixed bugs related to Python compatibility
• Fixed read counts from tools in final results table
• Updated models and provide additional models for feature subsets
• Cleaned code and made it more robust
• Updated error handling
• Cleaned up Dockerfile and made versioning more strict

1.3.6

• Add support for Singularity
• Update example output files
• Make Dockerfile more flexible
• Update README

1.3.5

• used a breakpoint-specific identifier (BPID) for joined annotation and in output
• output changes:
  • new output file names
  • separate output files for predicted fusions .pred.csv and all candidates .all.csv
  • new output format including column BPID
  • fixed content of columns <tool>_detected, tool_count, and tool_frac
• retrained model on new output column format
• cleaned up R code and updated R dependencies
• added Docker example scripts with test data and run_test.sh script
• added support for INI and JSON config files and make them more user-friendly
• fixed several bugs in input file/folder parsing

1.3.4

• updated prediction models EF_full (default) and EF_full_rep (with replicates).
• Fixed bug in fastqc parsing from different sources