CageWater: Accurately predicting water properties at rocket speed
Lakshmanji Verma and Ken Dill
Laufer Center for Physical and Quantitative Biology, Stony Brook University, Stony Brook, NY , USA
Water is everywhere, and life depends on it—we are about 70% water. Despite its ubiquity, water remains poorly understood. It behaves unlike other liquids: it has a density maximum, an unusually high heat capacity, and other anomalies that make it both scientifically fascinating and challenging to model. Atomistic simulations have been the only reliable way to predict water’s properties, but they are computationally expensive and typically reliable only near ambient conditions. CageWater overcomes these limitations with a statistical-mechanics–based analytical model. It predicts water properties roughly 1000× faster than explicit simulations, works across a wide range of temperatures and pressures, and resolves long-standing debates about supercooled water. It could also enable faster, more reliable simulations of biomolecules (proteins, DNA, drugs, etc.) under diverse conditions, accelerating drug discovery.