Brhanu F. Znabu

The paper benchmarks four splice-disruption predictors using experimental data, finding Pangolin performs best and identifying a shared blind spot in exon interiors.

NativeReady provides a benchmark and sequence-based model to predict native mass spectrometry suitability, achieving an AUC of 0.869 with cluster-aware cross-validation.

The paper develops an interpretable deep learning framework for multi-omics survival prediction in HCC, achieving better performance than existing models and identifying biologically relevant features.