AI-based methods for biomolecular structure modeling for Cryo-EM

Current Opinion in Structural Biology, Journal Year: 2025, Volume and Issue: 90, P. 102989 - 102989

Published: Jan. 27, 2025

Language: Английский

---

Mettl15-Mettl17 modulates the transition from early to late pre-mitoribosome

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 4, 2025

The assembly of the mitoribosomal small subunit involves folding and modification rRNA, its association with proteins. This process is assisted by a dynamic network factors. Conserved methyltransferases Mettl15 Mettl17 act on solvent-exposed surface rRNA. Binding associated early stage, whereas involved in late but mechanism transition between two was unclear. Here, we integrate structural data from Trypanosoma brucei mammalian homologs molecular dynamics simulations. We reveal how interplay intermediate steps links distinct stages assembly. analysis suggests model wherein acts as platform for recruitment. Subsequent release allows conformational change substrate recognition. Upon methylation, adopts loosely bound state which ultimately leads to replacement initiation factors, concluding Together, our results indicate that factors cooperate regulate biogenesis process, present resource understanding adaptations mitoribosome.

Language: Английский

---

Advancing structure modeling from cryo-EM maps with deep learning

Biochemical Society Transactions, Journal Year: 2025, Volume and Issue: 53(01)

Published: Feb. 7, 2025

Cryo-electron microscopy (cryo-EM) has revolutionized structural biology by enabling the determination of biomolecular structures that are challenging to resolve using conventional methods. Interpreting a cryo-EM map requires accurate modeling underlying biomolecules. Here, we concisely discuss evolution and current state automatic structure from density maps. We classify methods into two categories: de novo high-resolution maps (better than 5 Å) model fitting individual component proteins at lower resolution (worse Å). Special attention is given role deep learning in process, highlighting how AI-driven approaches transformative modeling. conclude discussing future directions field.

Language: Английский

---

Fitting Atomic Structures into Cryo-EM Maps by Coupling Deep Learning-Enhanced Map Processing with Global-Local Optimization

Journal of Chemical Information and Modeling, Journal Year: 2025, Volume and Issue: unknown

Published: March 28, 2025

With the breakthroughs in protein structure prediction technology, constructing atomic structures from cryo-electron microscopy (cryo-EM) density maps through structural fitting has become increasingly critical. However, accuracy of constructed models heavily relies on precision structure-to-map fitting. In this study, we introduce DEMO-EMfit, a progressive method that integrates deep learning-based backbone map extraction with global-local pose search to fit into maps. DEMO-EMfit was extensively evaluated benchmark data set comprising both tomography (cryo-ET) and cryo-EM nucleic acid complexes. The results demonstrate outperforms state-of-the-art approaches, offering an efficient accurate tool for

Language: Английский

---

DEMO-EMol: modeling protein-nucleic acid complex structures from cryo-EM maps by coupling chain assembly with map segmentation

Nucleic Acids Research, Journal Year: 2025, Volume and Issue: unknown

Published: May 14, 2025

Abstract Atomic structure modeling is a crucial step in determining the structures of protein complexes using cryo-electron microscopy (cryo-EM). This work introduces DEMO-EMol, an improved server that integrates deep learning-based map segmentation and chain fitting to accurately assemble protein–nucleic acid (NA) complex from cryo-EM density maps. Starting independently modeled structures, DEMO-EMol first segments NA regions learning. The overall then assembled by models into their respective segmented maps, followed domain-level optimization for chains. output includes final model along with residue-level quality assessments. was evaluated on comprehensive benchmark set maps resolutions ranging 1.96 12.77 Å, results demonstrated its superior performance over state-of-the-art methods both protein-NA protein–protein modeling. web freely accessible at https://zhanggroup.org/DEMO-EMol/.

Language: Английский

---

Improved automated model building for cryo-EM maps using CryFold

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 15, 2024

Constructing atomic models from cryogenic electron microscopy (cryo-EM) density maps is essential for interpreting molecular mechanisms. In this study, we present CryFold, an approach de novo model building in cryo-EM that leverages recent advancements AlphaFold2 1 to improve the state-of-the-art method ModelAngelo 2 . To incorporate map information, CryFold replaces global attention mechanism local attention, further enhanced by a novel 3D rotary position embedding. It produces more complete models, accelerates modeling, and reduces resolution requirement. Applying new results accurate differentiation between paralog sequences noisy regions, detection of previously uncharacterized proteins with unknown functions, precise compartmentalisation isolation non-protein components, improved modeling conformational changes. particular case, 104-protein complex was modeled within only 5.6 hours, minor change single protein domain detected at periphery when two different were compared. stands as currently available structure determination. open-source software https://github.com/SBQ-1999/CryFold

Language: Английский

---