RFcaller

RFcaller detects somatic mutations in tumor-normal paired samples using a machine learning-based pipeline to identify substitutions and insertions/deletions from whole-genome and whole-exome sequencing.

Key Features:

• Machine Learning Integration: RFcaller leverages machine learning algorithms to improve identification of somatic mutations from sequencing data.
• Read-level Features: The pipeline uses read-level features extracted from tumor-normal paired samples to inform variant calling.
• Efficiency and Low Computational Demand: RFcaller operates with minimal computing resources compared to traditional pipelines.
• High Accuracy and Validation: It detects substitutions and insertions/deletions from whole-genome and whole-exome data with validation reported against deep sequencing and Sanger sequencing.
• Detection of Missed Driver Mutations: RFcaller identifies mutations in driver genes that may be overlooked by other approaches.

Scientific Applications:

• Cancer Genomics Research: RFcaller provides high-confidence somatic mutation calls to support characterization of cancer genomes and identification of driver alterations.
• Clinical Practice Integration: Its precision in detecting substitutions and insertions/deletions from WGS/WES data supports applications where accurate somatic mutation detection is required for diagnosis or treatment planning.

Methodology:

RFcaller employs a machine learning-based pipeline that utilizes read-level features from tumor-normal paired samples.