RNAIndel

RNAIndel identifies and classifies somatic insertions and deletions (indels) from tumor RNA-sequencing (RNA-Seq) data to distinguish somatic, germline, and artifact events in expressed tumor transcriptomes.

Key Features:

• Indel Classification: Classifies indels into somatic, germline, and artifact categories based on their origin and characteristics within tumor transcriptomes.
• Postprocessing Capability: Can serve as a postprocessor to classify indels identified by other variant callers.
• Genome Build Support: Supports both GRCh38 and GRCh37 genome builds.
• Artifact-aware Detection: Accounts for artifacts introduced during PCR-based library preparation when distinguishing true variants from artifacts.
• Tumor-only Operation: Detects expressed somatic indels from tumor RNA-Seq data in the absence of matched normal RNA-Seq.

Scientific Applications:

• Somatic Indel Prediction: Predicts somatic indels across various cancer types, including pediatric and adult cancers.
• Subclonal Driver Indel Recovery: Recovers subclonal driver indels with variant allele frequencies (VAF) in the 0.01 to 0.15 range.
• Performance Superiority: Achieves sensitivity between 88% and 100% while producing 14 times fewer false positives than best-practice RNA-Seq variant calling methods.

Methodology:

Employs a machine-learning framework that uses features derived from the sequence context surrounding indels and their biological effects to differentiate somatic events from artifacts and germline variations.