Masayuki Nagai

Computational Biologist, Ph.D. student, Cold Spring Harbor Laboratory, New York

About Me

I’m a third-year Ph.D. student in the Koo Lab at Cold Spring Harbor Laboratory.

My research focuses on developing deep learning methods that improve the generalization and interpretability of biological models, with the goal of advancing our understanding of molecular and cellular mechanisms across genetics, epigenetics, and proteomics. By connecting predictive models with mechanistic insight, I aim to enable more informative characterization of disease processes and support the design of more effective therapeutics. I am also committed to the responsible advancement of biological AI models, with particular attention to biosecurity and AI alignment to mitigate dual-use risks and ensure safe deployment.

Outside of research, I’m passionate about tennis and climbing mountains!

Interests

  • Continual Learning: Refining causal understanding through perturbation data integration
  • Active Learning: Identifying the most informative experiments to guide research
  • Sequence Design: Engineering sequences with desired functional properties
  • Biosecurity: Preventing misuse and mitigating dual-use risks of biological AI models
  • Mechanistic Interpretability: Understanding the inner workings of deep learning models

News

  • Feb 2026 Released our review/perspective Toward Interpretable and Generalizable AI in Regulatory Genomics as a preprint on arXiv (co-first author). See a quick overview on X.
  • Aug 2025 Gave a talk on causal refinement for genomic deep learning models via Continual Learning at MLCB 2025 (YouTube).
  • Aug 2024 Joined Peter Koo's group at Cold Spring Harbor Laboratory.
  • Aug 2023 Started Ph.D. in the School of Biological Sciences, Cold Spring Harbor Laboratory.