FilterFFPE

FilterFFPE removes artificial chimeric reads from formalin-fixed paraffin-embedded (FFPE) next-generation sequencing (NGS) data to reduce false-positive split-read support and improve structural variant (SV) calling accuracy.

Key Features:

• Targeted Filtration: Targets and removes artificial chimeric reads characterized by erroneous split-read support while preserving genuine chimeric reads.
• Input Requirements: Accepts an indexed BAM file from FFPE samples as input.
• Complementary Tools: Developed alongside SimFFPE, an R package that simulates FFPE NGS data by generating mixtures of artifact chimeric reads and normal reads for evaluation.
• Performance Evaluation: Assessed with Delly, Lumpy, and Manta on simulated and real FFPE samples, improving mean positive predictive value (PPV) from 0.27 to 0.48 in simulated samples and from 0.11 to 0.27 in real samples while maintaining or slightly enhancing sensitivity.

Scientific Applications:

• FFPE structural variant detection: Improves the reliability of SV detection in FFPE-derived genomic studies, including cancer research, by reducing false-positive SV calls caused by FFPE-induced artifact chimeric reads.

Methodology:

Processes an indexed BAM to identify and remove artificial chimeric split reads while preserving genuine chimeric reads; performance was evaluated by running Delly, Lumpy, and Manta on simulated datasets generated with the SimFFPE R package and on real FFPE samples.