Computer-Aided Drug Discovery for Undruggable Targets DOI
Qi Sun, Hanping Wang, Juan Xie

и другие.

Chemical Reviews, Год журнала: 2025, Номер unknown

Опубликована: Май 27, 2025

Undruggable targets are those of therapeutical significance but challenging for conventional drug design approaches. Such often exhibit unique features, including highly dynamic structures, a lack well-defined ligand-binding pockets, the presence conserved active sites, and functional modulation by protein-protein interactions. Recent advances in computational simulations artificial intelligence have revolutionized landscape, giving rise to innovative strategies overcoming these obstacles. In this review, we highlight latest progress approaches against undruggable targets, present several successful case studies, discuss remaining challenges future directions. Special emphasis is placed on four primary target categories: intrinsically disordered proteins, protein allosteric regulation, interactions, degradation, along with discussion emerging types. We also examine how AI-driven methodologies transformed field, from applications protein-ligand complex structure prediction virtual screening de novo ligand generation targets. Integration methods experimental techniques expected bring further breakthroughs overcome hurdles As field continues evolve, advancements hold great promise expand druggable space, offering new therapeutic opportunities previously untreatable diseases.

Язык: Английский

Deciphering driving forces of biomolecular phase separation from simulations DOI Creative Commons
Lars V. Schäfer, Lukas S. Stelzl

Current Opinion in Structural Biology, Год журнала: 2025, Номер 92, С. 103026 - 103026

Опубликована: Март 8, 2025

The formation and modulation of biomolecular condensates as well their structural dynamic properties are determined by an intricate interplay different driving forces, which down at the microscopic scale involve molecular interactions biological macromolecules surrounding solvent ions. Molecular simulations increasingly used to provide detailed insights into various processes thermodynamic forces play, thereby yielding mechanistic understanding aiding interpretation experiments level individual amino acid residues or even atoms. Here we summarize recent advances in field biocondensate with a focus on coarse-grained all-atom dynamics (MD) simulations. We highlight possible future challenges concerning computationally efficient physically accurate large complex systems.

Язык: Английский

Процитировано

1
Machine learning methods to study sequence–ensemble–function relationships in disordered proteins DOI Creative Commons
Sören von Bülow, Giulio Tesei, Kresten Lindorff‐Larsen

и другие.

Current Opinion in Structural Biology, Год журнала: 2025, Номер 92, С. 103028 - 103028

Опубликована: Март 12, 2025

Язык: Английский

Процитировано

1
Deep generative modeling of temperature-dependent structural ensembles of proteins DOI Creative Commons
Giacomo Janson, Alexander Jussupow, Michael Feig

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Март 13, 2025

Deep learning has revolutionized protein structure prediction, but capturing conformational ensembles and structural variability remains an open challenge. While molecular dynamics (MD) is the foundation method for simulating biomolecular dynamics, it computationally expensive. Recently, deep models trained on MD have made progress in generating at reduced cost. However, they remain limited modeling atomistic details and, crucially, incorporating effect of environmental factors. Here, we present aSAM (atomistic autoencoder model), a latent diffusion model to generate heavy atom ensembles. Unlike most methods, atoms space, greatly facilitating accurate sampling side chain backbone torsion angle distributions. Additionally, extended into first reported transferable generator conditioned temperature, named aSAMt. Trained large mdCATH dataset, aSAMt captures temperature-dependent ensemble properties demonstrates generalization beyond training temperatures. By comparing long simulations fast folding proteins, find that high-temperature enhances ability generators explore energy landscapes. Finally, also show our MD-based can already capture experimentally observed thermal behavior proteins. Our work step towards generalizable generation complement physics- based approaches.

Язык: Английский

Процитировано

1
Computer-Aided Drug Discovery for Undruggable Targets DOI
Qi Sun, Hanping Wang, Juan Xie

и другие.

Chemical Reviews, Год журнала: 2025, Номер unknown

Опубликована: Май 27, 2025

Undruggable targets are those of therapeutical significance but challenging for conventional drug design approaches. Such often exhibit unique features, including highly dynamic structures, a lack well-defined ligand-binding pockets, the presence conserved active sites, and functional modulation by protein-protein interactions. Recent advances in computational simulations artificial intelligence have revolutionized landscape, giving rise to innovative strategies overcoming these obstacles. In this review, we highlight latest progress approaches against undruggable targets, present several successful case studies, discuss remaining challenges future directions. Special emphasis is placed on four primary target categories: intrinsically disordered proteins, protein allosteric regulation, interactions, degradation, along with discussion emerging types. We also examine how AI-driven methodologies transformed field, from applications protein-ligand complex structure prediction virtual screening de novo ligand generation targets. Integration methods experimental techniques expected bring further breakthroughs overcome hurdles As field continues evolve, advancements hold great promise expand druggable space, offering new therapeutic opportunities previously untreatable diseases.

Язык: Английский

Процитировано

0