SimFFPE and FilterFFPE: improving structural variant calling in FFPE samples.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Next-generation sequencing (NGS) of formalin-fixed paraffin-embedded (FFPE) samples is prone to artifact chimeric reads.
• These artifacts are misinterpreted as structural variants (SVs), leading to a high rate of false positives.
• Existing tools lack specific methods for calling or filtering SVs in FFPE data.

Purpose of the Study:

• To develop computational tools for addressing artifact chimeric reads in FFPE sequencing data.
• To improve the accuracy of structural variant detection in FFPE samples.

Main Methods:

• Development of SimFFPE, an R package for simulating FFPE sequencing data with artifact chimeric reads.
• Development of FilterFFPE, an R package designed to filter out artifact chimeric reads from sequencing data.
• Evaluation of FilterFFPE's performance by integrating it with standard SV callers (Delly, Lumpy, Manta) on simulated and real FFPE datasets.

Main Results:

• SimFFPE effectively simulates characteristic artifact chimeric reads alongside normal reads.
• FilterFFPE significantly reduces false-positive structural variant calls.
• After FilterFFPE application, the mean positive predictive value for SV calling improved from 0.27 to 0.48 in simulated data and from 0.11 to 0.27 in real data, with maintained or slightly increased sensitivity.

Conclusions:

• The FilterFFPE package demonstrably enhances the performance of structural variant calling in FFPE samples.
• Validation on both simulated and real datasets confirms FilterFFPE's efficacy in improving SV detection accuracy.