The orthogonal tilt reconstruction method

Abstract

Generating reliable initial models for novel asymmetric molecules, particularly heterogeneous ones, remains a major challenge in cryo-electron microscopy. Geometric reconstruction methods, relying on the ability to tilt the microscope stage to obtain two or more views of each molecule, are arguably the most robust for these types of samples as they generate independent reconstructions for each characteristic view obtained. Random Conical Tilt (RCT) is the classic geometric reconstruction method. Pairs of images are collected at high tilt (around 50°) and 0°. The latter are used to sort the data into characteristic views of the molecule and the former are used for their reconstruction. RCT's greatest strength is its ability to generate structures regardless of the number of orientations adopted by the sample on the support. Its major drawback stems from the limited tilt of the microscope stage; this results in an incomplete sampling of the structure in Fourier space and artifacts in its real space representation. Orthogonal Tilt Reconstruction (OTR), a modification of this data collection strategy, results in fully sampled structures. It relies on collecting data at -45° and +45° and treating the tilt pairs as equivalent to the ideal 0°/90° that cannot be collected directly in the microscope. OTR requires a sample that adopts a large number of orientations on the support. Here, the RCT and OTR methods are reviewed and their performances with a biological test sample are compared. The steps required to apply OTR are also discussed.