Recent advances in structure determination of subcellular systems using cryo-electron microscopy (cryoEM) and tomography (cryoET) have enabled a more detailed understanding of their three-dimensional architecture, providing valuable insights into their functions. However, several limitations, ranging from sample preparation to data collection and analysis to converting data intro structural models still hinder the potential of these techniques to provide valuable biological information [1]. Some of these challenges can now be overcome by combining new methodological developments in cryoEM/ET with physics-based and AI-powered computational approaches.

A notable example are Molecular Dynamics (MD) simulations, which are increasingly being used to complement the information derived from cryoEM experiments, in particular to facilitate flexible structure refinement (especially for heterogeneous and non-isotropic resolution), and to extract thermodynamic and kinetic properties from cryoEM data [2]. At the same time, cryo-electron tomograms, where identification and annotation of biological entities is still a major challenge, can benefit from new automated and standardized methods for molecular assignment to enable reconstruction of complex cellular landscapes [3]. These macromolecular assemblies determined by cryoEM/ET can then be used as starting points for in silico MD simulations and ultimately to build a dynamic model of the entire cell [4-6]. Finally, recent Artificial Intelligence (AI) approaches can be used to propose in silico models of molecular assemblies not yet determined by  structural biology experimental approaches [7], which can be further refined into cryoEM/ET envelopes to populate complex cellular landscapes.

Exploring these exciting research directions will be possible only if new developments are pursued to integrate experimental data with state-of-the-art computational approaches through close collaborations between computational and experimental groups  [8, 9]. The goal of our workshop is to gather experts in cryoEM/ET experimental techniques and researchers developing and applying state-of-the-art computational modelling approaches. These include physics-based methodologies to complement cryoEM/ET data with dynamic information, such as MD simulations, as well as novel AI-based approaches for structure prediction from sequence (AlphaFold, OpenFold, RoseTTAFold), and image analysis techniques for cryoEM/ET data (CryoDRGN, IsoNET, …), in particular those aimed at characterizing the conformational heterogeneity of biological systems directly from the data.  The ultimate goal of our workshop is to connect researchers with complementary expertise to discuss recent developments, identify challenges and bottlenecks in their respective fields, and foster new collaborative strategies.

More specifically, this workshop has the following objectives:

· Present an overview of recent advances in both cryoEM/ET data acquisition and analysis  as well as computational modelling approaches that exploit cryoEM/ET data;

· Discuss open issues and identify challenges related to experimental techniques (sample preparation optimization; automated, high-throughput multigrid acquisitions of cryoEM/ET) as well as computational approaches (on the fly processing of SPA projects; generalized pipelines for automated subtomogram averaging and tomogram segmentation; modelisation of dynamic or unstructured biomolecules; adaptive creation and training of neural networks for the different tasks…)

 · Identify integrative experimental and computational approaches that can help overcome the current challenges and advance the field ;

  ·  Offer opportunities for students and early-career researchers to present their projects in a poster session and through contributed talks;

  · Promote networking between students, early-career and more experienced researchers;

 ·  Foster collaborations between computational and experimental groups at all stages of their career.