v1.1.0

What's Changed

New Features:

• Trio Analysis Pipeline: Introduced support for Trio pipelines using the --input_ped argument, enabling family-based variant analyses.
• Single Variant Analysis: Added --input_variant parameter to handle analysis of single variants.
• Phenopacket Input Support: Pipeline now accepts --input_phenopacket, alternative to --input_hpo.
• Impact Filtering: Added an option to ignore low-impact transcripts during variant analysis.
• Manifest Management: Implemented automatic publishing of manifest.json files including detailed version and dependency information.
• Model Interpretability: Integrated SHAP for model interpretability, enhancing the analysis transparency.

Improvements:

• Improved Accuracy:: Enhanced the accuracy in an internal benchmark.
• Pipeline Validation: Enhanced pipeline robustness through VCF validation before preprocessing.
• Input Parameter Validation: Validate the parameter validation through nf-schema plugin.
• Codebase Maintenance: Added comprehensive pre-commit hooks, linting checks, and continuous integration workflows.
• Avoid Redundant Calculation: Introduced a caching mechanism to avoid redundant calculation. The cache directory can be specified with --storedir parameter.

Bug Fixes:

• Resolved a bug that Variant IDs sometimes differ from the user input. f2f108f
• Resolved file path handling issues. b1417d3
• Corrected BED file position indexing from 1-based to 0-based. b1417d3
• Fixed Unicode handling issues in VCF normalization. 9cf9564

Updates:

• Upgraded dependency Scikit Learn to version >= 1.3.0.
• Updated HGMD version to 2024.1
• Updated HPO version to 2024-08-13

Miscellaneous:

• Removed integrated output file.
• Simplified extensive parameter handling with the PREPARE_DATA subworkflow.
• Separated VEP and R into dedicated Dockerfiles.

---

For detailed commit logs, refer to the project repository history.
Full Changelog: v1.0.0...v1.1.0