cryo-em computational ai-structure
Princeton University
Info
Assistant Professor at Princeton University. Developer of cryoDRGN, a deep learning-based method for reconstructing continuous conformational heterogeneity from single-particle cryo-EM data. CryoDRGN uses variational autoencoders to learn a continuous latent space of 3D structures from raw cryo-EM images, enabling visualization of protein dynamics without the need for discrete 3D classification.
Key Relationships
- Princeton University — Assistant Professor
- cryoDRGN — principal developer
- Pilar Cossio — Bayesian cryo-EM heterogeneity community (complementary approach)
- Sjors Scheres — RELION 3D classification comparison point
- Erice International School of Crystallography — 2026 lecturer