Advancing structure modeling from cryo-EM maps with deep learning.

• Tác giả: Daisuke Kihara, Shu Li, Genki Terashi, Zicong Zhang
• Ngôn ngữ: eng
• Ký hiệu phân loại: 968.91 *Zimbabwe
• Thông tin xuất bản: England : Biochemical Society transactions , 2025
• Bộ sưu tập: NCBI
• ID: 9560

Tóm tắt

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 of the structures of underlying biomolecules. Here, we concisely discuss the evolution and current state of automatic structure modeling from cryo-EM density maps. We classify modeling methods into two categories: de novo modeling methods from high-resolution maps (better than 5 Å) and methods that model by fitting individual structures of component proteins to maps at lower resolution (worse than 5 Å). Special attention is given to the role of deep learning in the modeling process, highlighting how AI-driven approaches are transformative in cryo-EM structure modeling. We conclude by discussing future directions in the field.