Computational biophysics articles from across Nature Portfolio

Adupa et al show how the anti-amyloidogenic molecular chaperone DNAJB6 adopts three conformational states that determine the accessibility of its substrate binding domain. In all states, interactions with HSP70 are shielded, suggesting that functional interactions only may occur upon substrate binding.

Multiple parameter estimation techniques are employed to empirically validate theoretical propositions regarding complex systems by discerning relevant free parameters from often scarce experimental data. In this tutorial, the authors provide a beginner’s guide to parameter estimation via adjoint optimization, and show its efficiency in prototypical problems across different fields of physics.

Work by Klyshko and Kim et al. lays the foundation for simulating pump-probe experiments and demonstrates how the dynamic behaviour of proteins extends to the crystal environment, emphasizing the need for an ensemble view in understanding functional motions.

Here the authors perform modelling to reveal that the timescale of actin-VASP interactions plays a critical role in actin ring formation and filament length determines droplet deformation in VASP droplets: predictions from the model were tested against VASP GAB mutant.

A fast and versatile three-dimensional cell-based model, called SimuCell3D, is developed for high-resolution simulations of large and complex biological tissues. SimuCell3D natively integrates intra- and extracellular entities, including extracellular matrix, nuclei and polarized cell surfaces.

Intrinsically disordered regions of proteins are prevalent across the kingdoms of life; however, biophysical characterization is expensive, requiring specialized expertise and equipment and time-consuming sample preparation. By combining simulations and deep learning, we have developed a method to predict their average ensemble properties directly from sequence.

Orderly or coherent multicellular flows are fundamental in biology, but their triggers are not understood. In epithelial tissues, the tug-of-war between cells is now shown to lead to intrinsic asymmetric distributions in cell polarities that drive such flows.

Two new methods enable the study of the mechanics of tissue development.

Recent improvements in de novo protein design are likely to support a broad range of applications, but larger complexes will be easier to create if a building block approach is adopted. Now protein filaments with tunable geometry can be made using assemblies that have both cyclic and superhelical symmetries aligned along the same axis.

CryoREAD automatically builds DNA–RNA atomic structure from cryo-EM maps. Backbone accuracy is typically >85% and the method is applicable for maps with RNA-only, DNA-only and DNA–RNA–protein complex structures. CryoREAD uses deep learning to identify structure information and subsequently construct the 3D structure of nucleic acids.