EphaGen

EphaGen estimates the probability that a next-generation sequencing (NGS) dataset will miss predefined clinically relevant variants by computing an arbiter parameter as a proxy for diagnostic sensitivity.

Key Features:

• Arbiter Parameter Estimation: EphaGen associates each NGS dataset with a single arbiter parameter that estimates the probability of missing any variant from a predefined spectrum, providing a measure of diagnostic sensitivity.
• Input Requirements: The method requires a BAM file and a pre-compiled VCF file listing targeted clinically relevant variants to assess detection performance.
• Comparison to Conventional Metrics: EphaGen evaluates dataset ability to detect predefined clinically relevant variants rather than relying solely on coverage depth and uniformity metrics.

Scientific Applications:

• Gene-specific clinical QC: Applied to sequencing data for Mendelian-disease genes including BRCA1/2 and CFTR to evaluate diagnostic sensitivity.
• Benchmarking across datasets: Validated on 14 runs, 43 blood samples, and 504 publicly available NGS datasets to assess robustness across runs and samples.

Methodology:

EphaGen estimates the likelihood of missing clinically relevant variants by associating an NGS dataset with an arbiter parameter that serves as a proxy for diagnostic sensitivity.