Jakobi, A. | Search bioRxiv

bioRxiv2020

Direct information estimation from cryo-EM Movies with CARYON

While cryo-EM with modern direct electron detectors has proven incredibly powerful, becoming a dominant technique in structural biology, the analysis of cryo-EM images is significantly complicated by their exceptionally low signal-to-noise ratio, limiting the accuracy of the parameterisation of the physical models required for successful classification and reconstruction. Micrographs from modern direct electron detectors are typically collected as dose-fractionated multi-frame movies to allow the recording of separated individual electron impacts. These detectors improve electron detection and allow for both inter-frame motion correction, and dose-dependent image filtering, lessening the overall impact of effects deleterious to the recovery of high-resolution information. In this study we measured the information content at each spatial frequency in cryo-EM movies as it accrues during the course of an exposure. We show that, as well as correction for motion and radiation damage, the use of the information within movies allows substantially improved direct estimation of the remaining key image parameters required for accurate 3D reconstruction: the image CTF and spectral SNR. We are developing "CARYON" {insert contrived acronym here}, as a LAFTER-family filter for cryo-EM movies based upon such measurements. CARYON is intended to provide the best parameter estimation and filtration possible for a single complete, or large sub-section from a, movie micrograph without the use of a previously refined density. We demonstrate its utility in both single-particle and tomographic cryo-EM data processing.

biophysics↗

bioRxiv2018

Thresholding of cryo-EM density maps by false discovery rate control

Cryo-EM now commonly generates close-to-atomic resolution as well as intermediate resolution maps from macromolecules observed in isolation and in situ. Interpreting these maps remains a challenging task due to poor signal in the highest resolution shells and the necessity to select a threshold for density analysis. In order to facilitate this process, we developed a statistical framework for the generation of confidence maps by multiple hypothesis testing and false discovery rate (FDR) control. In this way, 3D confidence maps contain separated signal from background noise in the form of local detection rates of EM density values. We demonstrate that confidence maps and FDR-based thresholding can be used for the interpretation of near-atomic resolution single-particle structures as well as lower resolution maps determined by subtomogram averaging. Confidence maps represent a conservative way of interpreting molecular structures due to minimized noise. At the same time they provide a detection error with respect to background noise, which is associated with the density and particularly beneficial for the interpretation of weaker cryo-EM densities in cases of conformational flexibility and lower occupancy of bound molecules and ions to the structure.

biophysics↗