Hanne S. Antila

Integrative structural biology seeks molecular descriptions consistent with complementary experimental measurements. A particular challenge arises for molecules whose broad conformational ensembles resist description by a few representative structures, such as natively unfolded proteins and fluid-phase biomembrane lipids. Although molecular simulations can generate such ensembles, their simultaneous agreement with a large number of experimental constraints is commonly prevented by model inaccuracies. Here we introduce autofocusing as an approach that integrates experimental information by automated finetuning of the physical model generating the ensemble. We demonstrate it for a fluid bilayer of the phospholipid POPC, for which no existing force field has simultaneously reproduced the complete set of 82 NMR C-H bond order parameters. Autofocusing brings all the 82 within experimental uncertainty while preserving agreement with independent small-angle X-ray scattering measurements and retaining or improving agreement with untargeted NMR relaxation rates and effective correlation times, thus supporting its description of membrane structure and dynamics beyond the calibration targets. The resulting simulations challenge a long-standing two-state interpretation of phosphatidylcholine headgroup structure, instead supporting a broad conformational ensemble poorly represented by crystal-derived conformers. These results establish autofocusing as a practical approach to constructing experimentally grounded conformational ensembles while retaining the computational tractability of classical molecular dynamics.

Ligand–protein interaction prediction is essential for structure-based drug discovery, but further development of machine-learning scoring functions is limited by a lack of high-quality training datasets. Here, we present DockBind, a curated dataset of modeled protein–ligand complexes paired with experimental binding affinity data, generated through a quality-controlled docking workflow. This workflow was developed based on an expanded docking benchmark, where we systematically evaluated free and template-based docking methods as well as ligand similarity and maximum common substructure (MCS) based pose filtering to generate and identify the most realistic poses. The resulting workflow was then incorporated into a pipeline integrating structural and binding affinity data from Binding MOAD, the Protein Data Bank, ChEMBL, and BindingDB, which resulted in 18,332 complexes across 612 protein targets, with high structural reliability and chemical diversity – the DockBind dataset – providing a valuable resource for training and benchmarking machine-learning scoring functions.

Owing to the increase of available computational capabilities and the potential for providing a more accurate description, polarizable molecular dynamics force fields are gaining popularity in modeling biomolecular systems. It is, however, crucial to evaluate how much precision is truly gained with increasing cost and complexity of the simulation. Here, we leverage the NMRlipids open collaboration and Databank to assess the performance of available polarizable lipid models─the CHARMM-Drude and the AMOEBA-based parameters─against high-fidelity experimental data and compare them to the top-performing nonpolarizable models. While some improvement in the description of ion binding to membranes is observed in the most recent CHARMM-Drude parameters, and the conformational dynamics of AMOEBA-based parameters are excellent, the best nonpolarizable models tend to outperform their polarizable counterparts for each property we explored. The identified shortcomings range from inaccuracies in describing the conformational space of lipids to excessively slow conformational dynamics. Our results provide valuable insights for the further refinement of polarizable lipid force fields and for selecting the best simulation parameters for specific applications.

Owing to the increase of available computational capabilities and the potential for providing more accurate description, polarizable molecular dynamics force fields are gaining popularity in modelling biomolecular systems. It is, however, crucial to evaluate how much precision is truly gained with the increased cost and complexity of the simulation. Here, we leverage the NMRlipids open collaboration and Databank to assess the performance of available polarizable lipid models—the CHARMM-Drude and the AMOEBA-based parameters—against high-fidelity experimental data and compare them to the top-performing non-polarizable models. While some improvement in the description of ion binding to membranes is observed in the most recent CHARMM-Drude parameters, and the conformational dynamics of AMOEBA-based parameters are excellent, the best non-polarizable models tend to outperform their polarizable counterparts for each property we explored. The identified shortcomings range from inaccuracies in describing the conformational space of lipids to excessively slow conformational dynamics. Our results provide valuable insights for further refinement of polarizable lipid force fields and for selecting the best simulation parameters for specific applications.

A self-consistent mean field direct simulation Monte Carlo (SCMFD) algorithm was recently proposed for simulating collision environments for a range of one-dimensional model systems. This work extends the one-dimensional SCMFD approach to three dimensions and introduces a variable time step (3D-vt-SCMFD), enabling the modeling of a considerably wider range of different collision environments. We demonstrate the performance of the augmented method by modeling a varied set of test systems: ideal gas mixtures, Poiseuille flow of argon, and expansion of gas into high vacuum. For the gas mixtures, the 3D-vt-SCMFD method reproduces the properties (mean free path, mean free time, collision frequency, and temperature) in excellent agreement with theoretical predictions. From the Poiseuille flow simulations, we extract flow profiles that agree with the solution to the Navier–Stokes equations in the high-density limit and resemble free molecular flow at low densities, as expected. The measured viscosity from 3D-vt-SCMF is ∼15% lower than the theoretical prediction from Chapman–Enskog theory. The expansion of gas into vacuum is examined in the effusive regime and at the hydrodynamic limit. In both cases, 3D-vt-SCMDF simulations produce gas beam density, velocity, and temperature profiles in excellent agreement with analytical models. In summary, our tests show that 3D-vt-SCMFD is robust and computationally efficient, while also illustrating the diversity of systems the SCMFD model can be successfully applied to.

Molecular dynamics (MD) simulations are widely used to monitor time-resolved motions of biomacromolecules, although it often remains unknown how closely the conformational dynamics correspond to those occurring in real life. Here, we used a large set of open-access MD trajectories of phosphatidylcholine (PC) lipid bilayers to benchmark the conformational dynamics in several contemporary MD models (force fields) against nuclear magnetic resonance (NMR) data available in the literature: effective correlation times and spin–lattice relaxation rates. We found none of the tested MD models to fully reproduce the conformational dynamics. That said, the dynamics in CHARMM36 and Slipids are more realistic than in the Amber Lipid14, OPLS-based MacRog, and GROMOS-based Berger force fields, whose sampling of the glycerol backbone conformations is too slow. The performance of CHARMM36 persists when cholesterol is added to the bilayer, and when the hydration level is reduced. However, for conformational dynamics of the PC headgroup, both with and without cholesterol, Slipids provides the most realistic description because CHARMM36 overestimates the relative weight of ∼1 ns processes in the headgroup dynamics. We stress that not a single new simulation was run for the present work. This demonstrates the worth of open-access MD trajectory databanks for the indispensable step of any serious MD study: benchmarking the available force fields. We believe this proof of principle will inspire other novel applications of MD trajectory databanks and thus aid in developing biomolecular MD simulations into a true computational microscope—not only for lipid membranes but for all biomacromolecular systems.

Molecular dynamics (MD) simulations are widely used to monitor time-resolved motions of biomacromolecules, although it often remains unknown how closely the conformational dynamics correspond to those occurring in real life. Here, we used a large set of open-access MD trajectories of phosphatidylcholine (PC) lipid bilayers to benchmark the conformational dynamics in several contemporary MD models (force fields) against nuclear magnetic resonance (NMR) data available in the literature: effective correlation times and spin-lattice relaxation rates. We found none of the tested MD models to fully reproduce the conformational dynamics. That said, the dynamics in CHARMM36 and Slipids are more realistic than in the Amber Lipid14, OPLS-based MacRog, and GROMOS-based Berger force fields, whose sampling of the glycerol backbone conformations is too slow. The performance of CHARMM36 persists when cholesterol is added to the bilayer, and when the hydration level is reduced. However, for conformational dynamics of the PC headgroup, both with and without cholesterol, Slipids provides the most realistic description, because CHARMM36 overestimates the relative weight of ~1-ns processes in the headgroup dynamics. We stress that not a single new simulation was run for the present work. This demonstrates the worth of open-access MD trajectory databanks for the indispensable step of any serious MD study: Benchmarking the available force fields. We believe this proof of principle will inspire other novel applications of MD trajectory databanks, and thus aid in developing biomolecular MD simulations into a true computational microscope—not only for lipid membranes, but for all biomacromolecular systems.