I'm a computational biophysicist studying how proteins fold and how tissues flow. I received my Ph.D. in computational biology and bioinformatics from Yale University in 2024, advised by Corey O'Hern. My doctoral work was equal parts protein structural informatics and soft matter physics. Since 2024, I have been a postdoctoral researcher in the Syracuse Physics Department, working with M. Lisa Manning. In collaboration with Alessandro Mongera, we have developed new discrete element models for dynamic and sparse tissues under tension. See below for highlights of my work and my CV.
How much information is needed to fold a protein? We find that surprisingly, a simple binary representation of whether each residue is either buried in the core or exposed on the surface is enough information to discriminate the correct protein fold from incorrect folds. Moreover, we find that residue burial was the most efficient encoding of a protein's fold, needing only 0.37 bits per residue; more efficient than contact maps, secondary structure, hydrogen bonding satisfaction and even machine learned embedding like Foldseek 3Di. Check out our preprint linked below.
Protein cores all pack to the same density (φ ≈ 0.55), but there has been no physical explanation for why. We show this value marks the onset of a jamming transition: as hydrophobic interactions drive collapse, the core undergoes a floppy-to-rigid transition with the same power-law scaling hallmarks seen in jammed particulate systems. To demonstrate this, I built a new all-atom protein model from scratch — using only hard-core repulsion, bond geometry, and weak hydrophobic attraction — that achieves near-zero Ramachandran and side-chain dihedral outliers without explicit dihedral restraints, staying within ~1 Å RMSD of crystal structures at jamming onset and refolding from partially unfolded states. Check out our paper published in PRX Life linked below.
Embryonic mesenchymal tissues are sparse, under tension, and flow like a fluid — a combination that naively should be unstable. We developed a new particle-based interaction model with hysteretic sticking and stochastic bond kinetics to show that contact inhibition of locomotion (CIL) resolves this paradox, generating a uniform, tensioned and fluid material. Check out our preprint linked below.
Protein cores pack to the same density as jammed repulsive amino acids — but proteins collapse via attraction, not compression. Is the correspondence deep or coincidental? I developed simulations of attractive and repulsive bead-spring polymers and disconnected disks to better understand the similarities between polymer collapse and compressed packings. See our work in Phys. Rev. E. linked below.