sfragkoul/synth4bench: synth4bench Release

🚀 First Public Release We are happy to announce the first public release of synth4bench, a bioinformatics framework for synthetic genomics data generation and benchmarking of tumor-only somatic variant calling algorithms. synth4bench enables systematic evaluation of variant caller performance using fully controlled synthetic NGS datasets with known ground truth, with particular focus on low allele frequency variants. ✨ Features Synthetic NGS data generation using the NEAT simulator Benchmarking pipeline for major somatic variant callers (Mutect2, FreeBayes, VarDict, VarScan2, LoFreq) Independent analyses for true SNVs, noise SNVs and indels Fully automated workflow (single command execution) Configurable experimental setup via parameters.yaml Reproducible environment and openly available datasets 📦 Included Data generation scripts Variant calling benchmarking workflow Downstream analysis and plotting tools Conda environment setup Documentation and user guide 📄 Citation If you use synth4bench, please cite: Fragkouli S-C. et al. Synth4bench: Synthetic Data Generation for Benchmarking Tumor-Only Somatic Variant Calling Algorithms. bioRxiv (2025) https://doi.org/10.1101/2024.03.07.582313 🤝 Contributions Feedback and contributions are welcome via GitHub issues.