EBCall

EBCall applies an empirical Bayesian framework to detect somatic mutations, including insertions and deletions (InDels), from high-throughput sequencing data to identify true somatic variants under low sequencing depth or low tumor content.

Key Features:

• Empirical Bayesian Framework: EBCall employs an empirical Bayesian statistical model to distinguish true somatic mutations from sequencing errors.
• Utilization of Multiple Non-Paired Normal Samples: It incorporates sequencing data from multiple non-paired normal samples to estimate sequencing error distributions and inform model parameters.
• Performance at Low Depths and Allele Frequencies: The method detects mutations at moderate and low allele frequencies (≤ 10%) and under low sequencing depths, enabling identification of variants in minor tumor subpopulations.
• Validation on Whole-Exome Sequencing Across Cancer Types: EBCall has been validated using whole-exome sequencing data from diverse cancer types.

Scientific Applications:

• Somatic mutation discovery in cancer genomics: Accurate calling of somatic single-nucleotide variants and InDels from high-throughput sequencing data across cancer genomes.
• Tumor heterogeneity and subclone detection: Detection of low-frequency variants enables exploration of tumor heterogeneity and minor tumor subpopulations.
• Support for precision oncology research: High-confidence detection of low-allele-frequency mutations can inform studies that contribute to personalized treatment strategies.

Methodology:

EBCall statistically evaluates differences in observed allele frequencies between tumor samples and paired normal samples and integrates sequencing data from multiple non-paired normal samples into an empirical Bayesian framework to model sequencing errors and estimate model parameters.