Computer simulations of proteins in water are currently too unreliable to predict how the amyloid beta peptide misfolds in Alzheimer’s disease, and a UK team is rebuilding the core mathematical model from scratch to fix this. The problem lies in the “force fields”—the equations that govern how atoms interact in a simulation. Current versions add or remove vague energy terms to patch errors, but this has failed to produce trustworthy results. The team proposes FFLUX, a force field that uses machine learning to calculate how each atom’s energy and charge distribution shifts with its neighbours, capturing polarisation and charge transfer in one clean scheme. This avoids the rigid approximations of existing models. If successful, FFLUX could make molecular dynamics simulations a reliable complement to lab experiments. This matters for drug design and understanding diseases like Alzheimer’s, where protein misfolding is central. The work is fundamental science—it overhauls the computational engine rather than targeting a specific application—but past improvements in simulation accuracy have directly accelerated pharmaceutical development. The team will release the software as DL_FFLUX, opening the tool to a wider community.
View original technical description
The literature shows that there is a serious problem at the heart of biomolecular simulation: it cannot reliably predict the structure and dynamics of proteins in aqueous solution. Thus, molecular dynamics simulation cannot usefully complement experiment while studying the onset of Alzheimer's disease. Popular force fields do not reliably reproduce the misfolding and aggregation of the intrinsically disordered amyloid beta peptide. Remedies usually add or take away ill-defined energy terms or repeatedly re-parameterise. Because this strategy has not solved the problem, force field architecture needs to be overhauled. We propose FFLUX, a truly novel force field, that is much closer to the underlying quantum reality and one that "sees the electrons". FFLUX exploits a parameter-free definition of an atom inside a system. Using machine learning FFLUX learns how atomic energies, charges and multipole moments vary with the surrounding atoms' geometry. As such it captures all polarisation and many-body effects, as well as charge transfer, in one streamlined scheme. The approach avoids perturbation theory and thus benefits from a clear treatment of short-range interactions. Moreover, FFLUX breaks free from the rigid-body constraints of advanced polarisable force fields. The well-defined atom at the heart of FFLUX enables physics-based machine learning. It uses kriging instead of neural nets, thereby reducing the training data size. Our careful work plan is rooted in amino acids and water clusters, and scaled up to the solvated amyloid beta peptide via a sequence of increasingly relevant systems, both in gas-phase and in water. We will introduce more sophisticated machine learning and implement state-of-art parallellisation on CPUs, GPUs and FPGAs, thereby offering a new user community the program DL_FFLUX. Bringing about a step change in biomolecular simulation is a huge task, even for this type of grant, but feasible as evidenced by proofs-of-concept from our group.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know