bims-lances 2024-06-09 papers

Issue of 2024‒06‒09
two papers selected by
James M. Krieger, National Centre for Biotechnology

Bioinformatics. 2024 Jun 05. pii: btae345. [Epub ahead of print]

CLEAPA: a framework for exploring the conformational landscape of Cryo-EM using Energy-Aware pathfinding algorithm.

MOTIVATION: Cryo-electron microscopy (cryo-EM) is a powerful technique for studying macromolecules and holds the potential for identifying kinetically preferred transition sequences between conformational states. Typically, these sequences are explored within two-dimensional energy landscapes. However, due to the complexity of biomolecules, representing conformational changes in two dimensions can be challenging. Recent advancements in reconstruction models have successfully extracted structural heterogeneity from cryo-EM images using higher-dimension latent space. Nonetheless, creating high-dimensional conformational landscapes in the latent space and then searching for preferred paths continues to be a formidable task.RESULTS: This study introduces an innovative framework for identifying preferred trajectories within high-dimensional conformational landscapes. Our method encompasses the search for the minimum energy path in the graph, where edge weights are determined based on the energy estimation at each node using local density. The effectiveness of this approach is demonstrated by identifying accurate transition states in both synthetic and real-world datasets featuring continuous conformational changes.
AVAILABILITY: The CLEAPA package is available at https://github.com/tengyulin/energy_aware_pathfinding/.
SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

DOI: https://doi.org/10.1093/bioinformatics/btae345

J Comput Biophys Chem. 2023 Jun;22(4): 489-501

Assessments of Variational Autoencoder in Protein Conformation Exploration.

Sian Xiao, Zilin Song, Hao Tian, Peng Tao.

Molecular dynamics (MD) simulations have been extensively used to study protein dynamics and subsequently functions. However, MD simulations are often insufficient to explore adequate conformational space for protein functions within reachable timescales. Accordingly, many enhanced sampling methods, including variational autoencoder (VAE) based methods, have been developed to address this issue. The purpose of this study is to evaluate the feasibility of using VAE to assist in the exploration of protein conformational landscapes. Using three modeling systems, we showed that VAE could capture high-level hidden information which distinguishes protein conformations. These models could also be used to generate new physically plausible protein conformations for direct sampling in favorable conformational spaces. We also found that VAE worked better in interpolation than extrapolation and increasing latent space dimension could lead to a trade-off between performances and complexities.

Keywords: Deep learning; Enhanced sampling; Molecular dynamics; Protein conformations; Variational autoencoder

DOI: https://doi.org/10.1142/s2737416523500217