James M. Holton

The paper introduces a method to optimize protein conformation connectivity in ensemble refinement, improving R-free and geometric accuracy in models of multiple conformations.

WaterFlow predicts ordered water molecule positions in protein structures with sub-angstrom accuracy, outperforming existing methods and improving drug design and crystallographic refinement.

The paper identifies density misfit barrier traps that limit macromolecular model accuracy and introduces a synthetic dataset to challenge and improve ensemble refinement algorithms.

The paper develops MultiXray, a Bayesian method for multi-state and multi-condition protein structure modeling using X-ray crystallography data, improving Rfree values in both simulated and experimental cases.

This paper shows that deep learning language models can generate functional artificial proteins across families, with activities similar to natural proteins and structures resembling natural folds.