FusionGeneEngine

Introduction

Fusion genes arising from chromosomal translocations, inversions, and deletions are among the most clinically actionable cancer alterations. BCR-ABL1 in CML, EML4-ALK in NSCLC, and TMPRSS2-ERG in prostate cancer have transformed targeted therapy. We present FusionGeneEngine, a pure-Python pipeline for fusion gene detection and functional characterization.

Methods

Read-Level Evidence Filtering

For each fusion candidate, evidence is aggregated from:

• Split reads: reads spanning the fusion breakpoint
• Discordant pairs: read pairs mapping to different genes
• Junction reads: reads with soft-clipped sequences matching the partner gene

Quality filters: minimum 3 spanning reads, allele frequency > 0.05, mapping quality > 20.

In-Frame Fusion Prediction

Exon boundary phases (0, 1, 2) are used to predict reading frame preservation. In-frame fusions (phase match) are prioritized as likely to produce functional chimeric proteins.

Oncogenic Scoring

Composite score integrating:

1. COSMIC cancer gene census membership (both partners)
2. Known fusion database match (20 canonical fusions: BCR-ABL1, EML4-ALK, TMPRSS2-ERG, etc.)
3. In-frame status
4. Expression level of chimeric transcript
5. Recurrence across samples

Precision/Recall Evaluation

Ground truth: 50 true fusions injected into synthetic data. Precision and recall computed at score threshold 5.0.

Results

• 200 fusion candidates evaluated
• 52 high-confidence fusions (score > 5.0)
• Precision=0.962, Recall=1.000, F1=0.980
• 17 in-frame fusions
• 14 matching known oncogenic database
• Top: BCR-ABL1 (score=9.8, CML driver, in-frame, COSMIC tier 1)

Conclusion

FusionGeneEngine provides a complete, executable fusion gene detection pipeline achieving near-perfect recall with high precision on synthetic RNA-seq data.

Code

https://github.com/junior1p/FusionGeneEngine

pip install numpy scipy pandas matplotlib
python fusion_gene_engine.py