Modeling and Simulation Group Meeting

Cryo-electron microscopy (Cryo-Em) is a popular experimental technique for imaging proteins relevant to structural biology.

Very roughly, this technique involves (i) preparing a frozen sheet of ice containing a scattered jumble of many protein molecules, and (ii) collecting 2d-images of those molecules, each oriented in a random fashion.

In this context, I'll review some of the computational details associated with the so-called 'single-particle reconstruction' pipeline, which aims to reconstitute a 3d-volume from an assortment of 2d-images.

Along the way, I'll describe some recent research, including various computational strategies for improving the accuracy, robustness and speed of `ab-initio' single-particle-reconstruction -- i.e., when you don't yet know what the true protein molecule should look like, and you don't have much data to work with.

Given time, I'd also like to explain some of the challenges associated with 'multi-particle reconstruction', which might be necessary when the ice-sheet contains multiple different proteins or floppy molecules in a variety of conformations.

Towards the end, I'll spend some time discussing some of the applications in this area, and the related challenges that present opportunities for new research.