Computational Approaches for Protein–DNA/RNA Complex Modeling for Cryo‐EM Maps

Abstract

Abstract Cryogenic electron microscopy (cryo‐EM) has become a key method in structural biology for determining macromolecular structures. Numerous computational tools have been developed to build atomic models from cryo‐EM density maps. However, relatively few tools are available for modeling protein–nucleic acid complexes. Here, we describe how to use two such methods developed by our group, ComplexModeler and CryoZeta, with a focus on modeling protein–nucleic acid complexes. Both tools are available through the EMSuite web server, a freely accessible platform that hosts multiple methods for cryo‐EM structure modeling and validation. ComplexModeler integrates DiffModeler and CryoREAD to construct protein–DNA/RNA complex structures at resolutions of up to 5 Å. DiffModeler employs a diffusion model for backbone tracing, followed by fitting AlphaFold2‐predicted protein structures into the traced backbone. CryoREAD identifies nucleotide components (phosphate, sugar, and base), constructs the backbone, assigns sequences, and builds full atomic models of DNA/RNA chains. CryoZeta uses a diffusion‐based generative model that integrates sequence‐based structure prediction with cryo‐EM density features to generate accurate models of proteins, nucleic acids, and their complexes. This article describes how to use these two tools on the EMSuite web server through two modeling examples. © 2026 The Author(s). Current Protocols published by Wiley Periodicals LLC. Basic Protocol 1: Protein–nucleic acid structure modeling using ComplexModeler on the EMSuite server Basic Protocol 2: Protein–nucleic acid structure modeling using CryoZeta on the EMSuite server