JZJessica Lu Zhou

Jessica Lu Zhou

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Jessica Lu Zhou has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Uncertainty-aware genomic deep learning with knowledge distillation

    by Jessica Zhou, Kaeli Rizzo, Trevor Christensen, Ziqi Tang, Peter K Koo

    Deep neural networks (DNNs) have advanced predictive modeling for regulatory genomics, but challenges remain in ensuring the reliability of their predictions and understanding the key factors behind their decision making. Here we introduce DEGU (Distilling Ensembles for Genomic Uncertainty-aware models), a method that integrates ensemble learning and knowledge distillation to improve the robustness and explainability of DNN predictions. DEGU distills the predictions of an ensemble of DNNs into a single model, capturing both the average of the ensemble’s predictions and the variability across them, with the latter representing epistemic (or model-based) uncertainty. DEGU also includes an optional auxiliary task to estimate aleatoric, or data-based, uncertainty by modeling variability across experimental replicates. By applying DEGU across various functional genomic prediction tasks, we demonstrate that DEGU-trained models inherit the performance benefits of ensembles in a single model, with improved generalization to out-of-distribution sequences and more consistent explanations of cis-regulatory mechanisms through attribution analysis. Moreover, DEGU-trained models provide calibrated uncertainty estimates, with conformal prediction offering coverage guarantees under minimal assumptions. Overall, DEGU paves the way for robust and trustworthy applications of deep learning in genomics research.


  • Open Access

    Designing DNA With Tunable Regulatory Activity Using Discrete Diffusion

    by Anirban Sarkar, Alejandra Duran, Yiyang Yu, Da-Wei Lin, Yijie Kang, Nirali Somia, Pablo Mantilla, Jessica Lu Zhou, Masayuki Nagai, Ziqi Tang, et al.

    Designing regulatory DNA with tunable and context-specific activity is a major goal in biotechnology and medicine. Deep generative models offer a promising route for sequence design, yet it remains unclear whether synthetic sequences faithfully recapitulate the motif organization and functional specificity of natural regulatory DNA. Here we present DNA Discrete Diffusion (D3), a generative model that designs regulatory DNA through an iterative nucleotide-substitution process. Across computational benchmarks, D3 improves regulatory sequence generation relative to matched diffusion baselines, producing sequences that more closely match target activity, activity distributions, and sequence composition. In K562 lentiMPRA experiments, D3-designed sequences retained measurable regulatory activity and more closely recapitulated the activity distribution of genomic regulatory sequences than matched diffusion baselines. D3 performs robustly with limited training data and generates sequences informative enough to improve predictive models when labeled data are scarce. When trained without activity labels on task-specific regulatory sequence sets, D3 learns frozen sequence representations that are predictive of enhancer activity and compare favorably to several off-the-shelf genomic language model embeddings. Analysis of the sampling process identifies reproducible phases of sequence exploration, compositional refinement, and motif-associated convergence, providing an interpretable view of how D3 constructs regulatory sequences. These results establish D3 as a practical framework for designing synthetic regulatory DNA and studying sequence features associated with context-specific regulatory activity.


  • Open Access

    Author Correction: Single-nucleus genomics in outbred rats with divergent cocaine addiction-like behaviors reveals changes in amygdala GABAergic inhibition

    by Jessica L. Zhou, Giordano de Guglielmo, Aaron J. Ho, Marsida Kallupi, Narayan Pokhrel, Hai-Ri Li, Apurva S. Chitre, Daniel Munro, Pejman Mohammadi, Lieselot L. G. Carrette, et al.


  • Open Access

    Single-nucleus genomics in outbred rats with divergent cocaine addiction-like behaviors reveals changes in amygdala GABAergic inhibition

    by Jessica L. Zhou, Giordano de Guglielmo, Aaron J. Ho, Marsida Kallupi, Narayan Pokhrel, Hai-Ri Li, Apurva S. Chitre, Daniel Munro, Pejman Mohammadi, Lieselot L. G. Carrette, et al.

    The amygdala processes positive and negative valence and contributes to addiction, but the cell-type-specific gene regulatory programs involved are unknown. We generated an atlas of single-nucleus gene expression and chromatin accessibility in the amygdala of outbred rats with high and low cocaine addiction-like behaviors following prolonged abstinence. Differentially expressed genes between the high and low groups were enriched for energy metabolism across cell types. Rats with high addiction index (AI) showed increased relapse-like behaviors and GABAergic transmission in the amygdala. Both phenotypes were reversed by pharmacological inhibition of the glyoxalase 1 enzyme, which metabolizes methylglyoxal—a GABAA receptor agonist produced by glycolysis. Differences in chromatin accessibility between high and low AI rats implicated pioneer transcription factors in the basic helix-loop-helix, FOX, SOX and activator protein 1 families. We observed opposite regulation of chromatin accessibility across many cell types. Most notably, excitatory neurons had greater accessibility in high AI rats and inhibitory neurons had greater accessibility in low AI rats.


  • Open Access

    Analysis of single-cell CRISPR perturbations indicates that enhancers act multiplicatively and provides limited evidence for epistatic-like interactions

    by Jessica Zhou, Karthik Guruvayurappan, Shushan Toneyan, Hsiuyi V. Chen, Aaron R. Chen, Peter Koo, Graham McVicker

    Summary: A single gene may have multiple enhancers, but how they work in concert to regulate transcription is poorly understood. To analyze enhancer interactions throughout the genome, we developed a generalized linear modeling framework, GLiMMIRS, for interrogating enhancer effects from single-cell CRISPR experiments. We applied GLiMMIRS to a published dataset and tested for interactions between 46,166 enhancer pairs and corresponding genes, including 264 ’high-confidence’ enhancer pairs. We found that enhancer effects combine multiplicatively but with limited evidence for further interactions. Only 31 enhancer pairs exhibited significant interactions (FDR < 0.1), of which none came from the high confidence subset and 20 were driven by outlier expression values. Additional analyses of a second CRISPR dataset and in silico enhancer perturbations with Enformer both support a multiplicative model of enhancer effects without interactions. Altogether, our results indicate that enhancer interactions are uncommon or have small effects that are difficult to detect. Highlights: Analysis of a large single-cell CRISPRi screen finds limited evidence for synergistic or redundant interactions between enhancers The collective action of multiple enhancers on a common target gene follows a multiplicative model of activity A new statistical framework for simulating and modeling data from single-cell CRISPRi screens


  • Open Access

    Cocaine addiction-like behaviors are associated with long-term changes in gene regulation, energy metabolism, and GABAergic inhibition within the amygdala

    by Jessica L. Zhou, Giordano de Guglielmo, Aaron J. Ho, Marsida Kallupi, Narayan Pokhrel, Hai-Ri Li, Apurva S. Chitre, Daniel Munro, Pejman Mohammadi, Lieselot Lg Carrette, et al.

    The amygdala processes positive and negative valence and contributes to the development of addiction, but the underlying cell type-specific gene regulatory programs are unknown. We generated an atlas of single nucleus gene expression and chromatin accessibility in the amygdala of outbred rats with low and high cocaine addiction-like behaviors following prolonged abstinence. Between rats with different addiction indexes, we identified thousands of cell type-specific differentially expressed genes enriched for energy metabolism-related pathways that are known to affect synaptic transmission and action potentials. Rats with high addiction-like behaviors showed enhanced GABAergic transmission in the amygdala, which, along with relapse-like behaviors, were reversed by inhibition of Glyoxalase 1, which metabolizes the GABA A receptor agonist methylglyoxal. Finally, we identified thousands of cell type-specific chromatin accessible sites and transcription factor (TF) motifs where accessibility was associated with addiction index, most notably at motifs for pioneer TFs in the Fox, Sox, helix-loop-helix, and AP1 families.


  • Automated analytical systems for drug development studies. I — A system for the determination of drug stability

    by K.P. Shah, J. Zhou, R. Lee, R.L. Schowen, R. Elsbernd, J.M. Ault, J.F. Stobaugh, M. Slavik, C.M. Riley