Fast and automated protein-DNA/RNA macromolecular complex modeling from cryo-EM maps

Abstract

Cryo-electron microscopy (cryo-EM) allows a macromolecular structure such as protein-DNA/RNA complexes to be reconstructed in a three-dimensional coulomb potential map. The structural information of these macromolecular complexes forms the foundation for understanding the molecular mechanism including many human diseases. However, the model building of large macromolecular complexes is often difficult and time-consuming. We recently developed DeepTracer-2.0, an artificial-intelligence-based pipeline that can build amino acid and nucleic acid backbones from a single cryo-EM map, and even predict the best-fitting residues according to the density of side chains. The experiments showed improved accuracy and efficiency when benchmarking the performance on independent experimental maps of protein-DNA/RNA complexes and demonstrated the promising future of macromolecular modeling from cryo-EM maps. Our method and pipeline could benefit researchers worldwide who work in molecular biomedicine and drug discovery, and substantially increase the throughput of the cryo-EM model building. The pipeline has been integrated into the web portal https://deeptracer.uw.edu/.

Structural basis of antiphage immunity generated by a prokaryotic Argonaute-associated SPARSA system

“…The backbone structures were traced in the cryo-EM maps using Deeptracer 60 , which guided the fitting of the Alfafold2 predicted structures into the density map. The model adjustment and building were then accomplished in Coot 61 and iteratively refined in real space using PHENIX 62 .…”

Automated model building and protein identification in cryo-EM maps

“…The sudden availability of atomic models for millions of proteins from protein structure prediction by AlphaFold2 15 , 16 has helped to guide and accelerate model building 17 . However, previous attempts to fully automate atomic modelling 18 – 24 or the identification of unknown proteins 25 – 27 have not become mainstream, although DeepTracer 21 , 24 and findMySequence 25 have gained some traction. However, atomic modelling remains a time-consuming and expert-dependent process in many structure determination projects.…”

Novel Artificial Intelligence-Based Approaches for Ab Initio Structure Determination and Atomic Model Building for Cryo-Electron Microscopy

“…While DeepTracer demonstrates more accurate Cα prediction than its counterparts [ 72 , 128 , 161 ], it is currently only applicable to maps with 5 Å resolution or higher. The latest version, DeepTracer-2.0 [ 61 ], also has the ability to model nucleic acids.…”

“…Recently, deep learning (DL) techniques have emerged as promising tools to tackle these aforementioned challenges because of their capability to learn complicated patterns and extract meaningful information from large and complex datasets. Although still in their infancy, various fully automated DL-based approaches have already proven useful for several image processing tasks, including particle picking [ 29 , 30 , 31 , 32 , 33 , 34 , 35 ], 3D reconstruction [ 36 , 37 , 38 , 39 , 40 , 41 , 42 , 43 , 44 , 45 , 46 , 47 ], local resolution estimation [ 48 , 49 ], and model building [ 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 59 , 60 , 61 , 62 , 63 , 64 , 65 , 66 , 67 ]. In this article, we explore the applications of new AI-based algorithms for two current bottlenecks of the cryo-EM image processing pipeline: ab initio reconstruction and de novo atomic model building.…”