Jessica Zhou

This paper introduces DEGU, a method that improves deep learning reliability in genomics by combining ensemble learning and knowledge distillation to capture both average predictions and uncertainty.

This paper introduces D3, a generative model for designing regulatory DNA with tunable activity, showing improved performance over baselines in sequence generation and activity prediction.

The paper analyzes single-cell CRISPR perturbations to show enhancers act multiplicatively with limited evidence for epistatic interactions.

The paper identifies cell type-specific gene regulatory changes in the amygdala of rats with varying cocaine addiction-like behaviors, linking them to energy metabolism, GABAergic inhibition, and chromatin accessibility.