Using a supporting geometry

In cryo-ET, objects of interest are often bound to or part of a larger supporting object such as a membrane, a vesicle or a filament. When this is the case, they often exhibit specific spatial relationships to the support.

This information, the position and orientation of particles relative to a supporting geometry, is often useful in subtomogram averaging experiments.

The process

The general process for making use of this spatial information is simple

  1. Generate a 3D model of the supporting geometry from tomogram annotations

  2. Seed particle positions and orientations relative to the supporting geometry


Providing accurate initial estimates for the positions and orientations of particles at the start of a subtomogram averaging experiment allows you to

  • limit angular searches appropriately

  • restrict shifts appropriately

Constraining these parameters ensures that particles do not ‘drift’ during iterative refinement procedures and get trapped in a local minimum, a common phenomenon when working with low SNR cryo-ET data.

A reduced search space also alleviates some of the computational burden of performing global searches.


Generation of these 3D models is usually semi-automated at best, as opposed to template matching which is fully automated. This can mean it takes significantly longer for large datasets.