Abstract

In this paper, several approaches to be used to accelerate algorithms for fitting an atomic structure into a given 3D density map determined by cryo-EM are discussed. Rotation and translation of the atomic structure to find similarity scores are used and implemented with discrete Fourier transforms. Several rotations can be combined into groups to accelerate processing. The finite resolution of experimental and simulated maps allows a reduction in the number of rotations and translations needed in order to estimate similarity-score values.

Highlights

  • The goal of many cryogenic electron microscopy studies is to obtain an atomic model or models corresponding to the data and the molecular complexes imaged

  • In cryogenic electron microscopy (cryo-electron microscopy (EM)) we image with electrons, so the map obtained is strictly an electron potential map

  • We introduce a mask to confine the signal compared to a fixed region around the search object, reducing the noise contributed by unmatched parts of the target map

Read more

Summary

Introduction

The goal of many cryogenic electron microscopy (cryo-EM) studies is to obtain an atomic model or models corresponding to the data and the molecular complexes imaged. Docking problems in electron microscopy (EM), i.e. locating a known density object optimally within a larger 3D density map, are similar to but different from molecularreplacement (MR) problems in X-ray crystallography. Our aim is to generate useful solutions fast enough that docking can be performed interactively by a user on a high-performance desktop workstation This will allow them to explore and guide model building using additional information and allow them to test their hypotheses in real time. Our second aim is to facilitate low-resolution model building for another case: where no known candidate homologous structures can be identified for a target map or some region of the map. The user could select atomic models to fit based on an appropriate size range This approach may allow us to limit the number of candidates to thousands.

Similarity of vectors for noisy data
Correlation map in 3D
Numerical implementation
Size of maps
Rotation and Fourier space
Cubic ROIs and an integer centre of rotation
False peaks
Smooth functions
KPÀ1mPÀ1
Real and complex DFTs
Run times
Conclusions
Findings
Atomic scattering factors
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call