subtomogram

Subtomogram is a three-dimensional image patch extracted from a cryo-electron tomography tomogram, containing a single instance of a macromolecular complex within its cellular context. Cryo-ET collects a tilt series of two-dimensional images and reconstructs a tomogram, in which multiple copies of the same particle may be present at different locations and orientations.

Extraction and processing: Positions of particles are identified by manual inspection or automated methods such as template matching or deep learning. Subvolumes surrounding these coordinates are extracted as subtomograms and then subjected to alignment to a common reference. Iterative refinement and averaging improve signal-to-noise and reveal structural details. Classification can separate different conformations or orientations before or during averaging.

Challenges and considerations: The limited tilt range produces a missing wedge that creates anisotropic resolution, complicating alignment and interpretation. Low signal-to-noise ratio, sample thickness, radiation damage, and potential bias from the reference pose present additional challenges. Strategies include masking, careful validation, independent half-sets, and incorporating prior information or probabilistic alignment.

Applications and output: Subtomogram averaging yields a high-SNR density map for the particle, which can be interpreted by fitting known structures or building models. It enables in situ structural studies of ribosomes, proteasomes, and other complexes inside cells or tissues, preserving native context.

Software and workflow: Popular tools include Dynamo, RELION with subtomogram averaging, EMAN2, and PyTom-based pipelines. The workflow typically ends with resolution estimation by Fourier shell correlation and validation of the structural model.