Twistcgp流程：一个用于转化肿瘤学中全面基因组分析的可移植开源工作流程

Comprehensive genomic profiling (CGP) panels are widely used in translational studies & clinical trials to support patient stratification, exploratory biomarker analysis, & retrospective molecular characterization. However, secondary analysis & reporting workflows are often delivered as tightly coupled, proprietary components of commercial assay ecosystems. This coupling limits transparency, reproducibility across sites, & the flexibility to adapt analytical methods as study designs & sequencing technologies evolve. We present twistcgp, an open-source bioinformatics workflow for analysis of data generated with the Twist Oncology DNA CGP Panel that targets 562 cancer-associated genes across 2.4 Mb with the ability to customize content This collaboration pairs a focused hybrid-capture CGP assay with a transparent, configurable, & portable analysis framework that operates independently of licensed software, proprietary execution environments, or fixed reporting layers. The twistcgp workflow is implemented in Nextflow & follows community best practices for reproducible workflow design, enabling standardized execution across local infrastructure, high-performance computing clusters, or cloud platforms. Starting from FASTQ input, the pipeline supports alignment, somatic SNV & indel detection, copy number variant analysis, tumor mutational burden (TMB) estimation following Friends of Cancer Research harmonization guidelines, microsatellite instability (MSI) classification using MSIsensor2, variant annotation via Ensembl VEP & the CIViC knowledgebase, & comprehensive quality control reporting. Pipeline performance was evaluated using Horizon Discovery, Seracare, and Twist reference standards sequenced with the Twist CGP Panel, including technical replicates. Across all samples, the workflow generated expected classes of somatic variation with standardized quality metrics suitable for cross-sample & cross-site comparison. Known variants in reference materials were detected at anticipated VAFs, & TMB & MSI estimates were consistent with expectations, derived directly from panel data with out reliance on proprietary analysis software. Execution performance was also assessed on the Seqera Platform to support scalable deployment. By decoupling CGP assay chemistry from analysis, twistcgp enables reproducible genomic analysis across distributed research environments. Together, the Twist Oncology DNA CGP Panel & twistcgp workflow demonstrate a practical approach to CGP that reduces analytical friction, supports methodological consistency, & enables flexible integration into translational oncology studies. Fusion detection is planned for a future release. The pipeline is freely available on GitHub under MIT license for research use only.