FusionAI

FusionAI predicts human fusion gene breakpoints from DNA sequences generated by next-generation sequencing by applying deep learning and genomic feature analysis to reduce false positives.

Key Features:

• Deep learning pipeline: FusionAI employs deep learning models to analyze DNA sequence context and predict precise fusion gene breakpoint locations.
• Genomic feature integration: The method leverages genomic features surrounding breakpoints to distinguish true fusion events from false positives.
• NGS data compatibility: The model operates on DNA sequence data derived from next-generation sequencing (NGS).

Scientific Applications:

• Cancer genomics: Improved breakpoint predictions facilitate identification of fusion oncogenes in cancer genomics studies.
• Oncogenic driver discovery: Aids discovery and prioritization of potential oncogenic fusion drivers.
• Disease mechanism investigation: Supports studies of molecular mechanisms in diseases involving fusion genes.

Methodology:

Deep learning models are trained on DNA sequence data from NGS and use analysis of genomic features surrounding candidate breakpoints to predict breakpoint positions and reduce false positives.