What Will It Take to Understand Biology with Cryo-Electron Tomography and Subtomogram Averaging? Moving from Breathtaking Renderings to Quantitative, Context-Aware Visual Proteomics
Date:
Invited seminar on the scientific and conceptual frontiers of cellular cryo-electron tomography and in situ visual proteomics.
Cryo-electron tomography (Cryo-ET) enables visualization of cellular structures in unprecedented detail and within their native context. In recent years, the subtomogram averaging has resolved select macromolecular complexes at subnanometer resolution within their native context. Yet for all the breathtaking renderings, the promise of visual proteomics remains unevenly realized. In our large-scale cryo-ET study of Chlamydomonas reinhardtii, we generated and shared ~1800 partially annotated tomograms. From this dataset, we obtained a handful of subnanometer resolution averages, including ribosomes, rubisco, clathrin, nucleosome, and mitochondrial respiratory chain. While these results demonstrate the potential of modern cryo-ET workflows, they also underscore a limitation: even with large-scale data and state-of-the-art computational methods, the number of distinct, interpretable structures remains small, limiting biological insights. To address this, our recent efforts have focused on developing workflows capable of quantitative classification and dissecting structural heterogeneity. This is critical, for example, when visualizing ribosome assembly intermediates within the nucleolus, where meaningful biological interpretation depends on statistically robust confidence measures of the structural state of individual particles. Similar challenges arise in the detection of rare or transient macromolecular complexes, where low abundance, flexibility, and compositional variability confound traditional particle identification and subtomogram averaging strategies. I will highlight these methodological and conceptual challenges, especially as they relate to detecting rare or transient macromolecular complexes within their native context and discuss how we might better define what it means to pursue visual proteomics at scale. Rather than chasing ever-higher resolution subtomogram averages of select macromolecular complexes, I propose that novel ab-initio approaches, spatial context, quantitative classification, and interpretable flexibility analysis may improve our ability to extract systems-level biological insight from complex cryo-ET data.
Invitation by: Florian Faessler