A comprehensive exploration of the druggable conformational space of protein kinases using AI-predicted structures DOI Creative Commons
Noah B. Herrington, Yan Chak Li, David Stein

и другие.

PLoS Computational Biology, Год журнала: 2024, Номер 20(7), С. e1012302 - e1012302

Опубликована: Июль 24, 2024

Protein kinase function and interactions with drugs are controlled in part by the movement of DFG ɑC-Helix motifs that related to catalytic activity kinase. Small molecule ligands elicit therapeutic effects distinct selectivity profiles residence times often depend on active or inactive conformation(s) they bind. Modern AI-based structural modeling methods have potential expand upon limited availability experimentally determined structures states. Here, we first explored conformational space kinases PDB models generated AlphaFold2 (AF2) ESMFold, two prominent protein structure prediction methods. Our investigation AF2’s ability explore diversity kinome at various multiple sequence alignment (MSA) depths showed a bias within predicted DFG-in conformations, particularly those motif, based their overabundance PDB. We demonstrate predicting using AF2 lower MSA these alternative conformations more extensively, including identifying previously unobserved for 398 kinases. Ligand enrichment analyses 23 that, average, docked distinguished between molecules decoys better than random (average AUC (avgAUC) 64.58), but select perform well (e.g., avgAUCs PTK2 JAK2 were 79.28 80.16, respectively). Further analysis explained ligand discrepancy low- high-performing as binding site occlusions would preclude docking. The overall results our suggested although uncharted regions exhibited scores suitable rational drug discovery, rigorous refinement is likely still necessary discovery campaigns.

Язык: Английский

Advancing structural biology through breakthroughs in AI DOI Open Access
Laksh Aithani, Eric Alcaide,

Sergey Bartunov

и другие.

Current Opinion in Structural Biology, Год журнала: 2023, Номер 80, С. 102601 - 102601

Опубликована: Май 12, 2023

Язык: Английский

Процитировано

22
Keeping pace with the explosive growth of chemical libraries with structure‐based virtual screening DOI Creative Commons

Jacqueline Kuan,

Mariia Radaeva,

Adeline Avenido

и другие.

Wiley Interdisciplinary Reviews Computational Molecular Science, Год журнала: 2023, Номер 13(6)

Опубликована: Июнь 20, 2023

Abstract Recent efforts to synthetically expand drug‐like chemical libraries have led the emergence of unprecedently large virtual databases. This surge make‐on‐demand molecular datasets has been received enthusiastically across drug discovery community as a new paradigm. In several recent studies, screening (VS) larger collections resulted in identification novel molecules with higher potency and specificity compared more conventional VS campaigns relying on smaller in‐stock libraries. These results inspired ultra‐large against various clinically relevant targets, including key proteins SARS‐CoV‐2 virus. As library sizes rapidly surpassed billion compounds mark, computational strategies emerged, shifting from docking fragment‐based machine learning‐accelerated methods. approaches significantly reduce demands screenings by lowering number explicitly docked onto target. Such already demonstrated promise evaluating tens billions at relatively low cost. Herein, we review advancements structure‐based methods for that practitioners adopted explore ever‐expanding universe. article is categorized under: Data Science > Databases Expert Systems Artificial Intelligence/Machine Learning Molecular Statistical Mechanics

Язык: Английский

Процитировано

20
Limitations of Protein Structure Prediction Algorithms in Therapeutic Protein Development DOI Creative Commons
Sarfaraz K. Niazi, Zamara Mariam, Rehan Zafar Paracha

и другие.

BioMedInformatics, Год журнала: 2024, Номер 4(1), С. 98 - 112

Опубликована: Янв. 8, 2024

The three-dimensional protein structure is pivotal in comprehending biological phenomena. It directly governs function and hence aids drug discovery. development of prediction algorithms, such as AlphaFold2, ESMFold, trRosetta, has given much hope expediting protein-based therapeutic Though no study reported a conclusive application these the efforts continue with optimism. We intended to test algorithms rank-ordering proteins for their instability during pre-translational modification stages, may be predicted according confidence by algorithms. selected molecules were based on harmonized category licensed proteins; out 204 products, 188 that not conjugated chosen analysis, resulting lack correlation between scores structural or properties. crucial note here predictive accuracy contingent upon presence known accessible database. Consequently, our conclusion emphasizes primarily replicate information derived from existing structures. While findings caution against relying discovery purposes, we acknowledge need nuanced interpretation. Considering limitations recognizing utility constrained scenarios where structures are available important. Hence, advised when applying characterize various attributes without support adequate information. worth noting two main AlfphaFold2 also showed 72% scores, pointing similar limitations. progress been made computational sciences, Levinthal paradox remains unsolved.

Язык: Английский

Процитировано

9
DiffBindFR: an SE(3) equivariant network for flexible protein–ligand docking DOI Creative Commons
Jintao Zhu, Zhonghui Gu, Jianfeng Pei

и другие.

Chemical Science, Год журнала: 2024, Номер 15(21), С. 7926 - 7942

Опубликована: Янв. 1, 2024

DiffBindFR, a diffusion model based flexible full-atom protein–ligand docking tool, demonstrates its superior and side-chain refinement accuracy with reliable physical plausibility.

Язык: Английский

Процитировано

9
A comprehensive exploration of the druggable conformational space of protein kinases using AI-predicted structures DOI Creative Commons
Noah B. Herrington, Yan Chak Li, David Stein

и другие.

PLoS Computational Biology, Год журнала: 2024, Номер 20(7), С. e1012302 - e1012302

Опубликована: Июль 24, 2024

Protein kinase function and interactions with drugs are controlled in part by the movement of DFG ɑC-Helix motifs that related to catalytic activity kinase. Small molecule ligands elicit therapeutic effects distinct selectivity profiles residence times often depend on active or inactive conformation(s) they bind. Modern AI-based structural modeling methods have potential expand upon limited availability experimentally determined structures states. Here, we first explored conformational space kinases PDB models generated AlphaFold2 (AF2) ESMFold, two prominent protein structure prediction methods. Our investigation AF2’s ability explore diversity kinome at various multiple sequence alignment (MSA) depths showed a bias within predicted DFG-in conformations, particularly those motif, based their overabundance PDB. We demonstrate predicting using AF2 lower MSA these alternative conformations more extensively, including identifying previously unobserved for 398 kinases. Ligand enrichment analyses 23 that, average, docked distinguished between molecules decoys better than random (average AUC (avgAUC) 64.58), but select perform well (e.g., avgAUCs PTK2 JAK2 were 79.28 80.16, respectively). Further analysis explained ligand discrepancy low- high-performing as binding site occlusions would preclude docking. The overall results our suggested although uncharted regions exhibited scores suitable rational drug discovery, rigorous refinement is likely still necessary discovery campaigns.

Язык: Английский

Процитировано

7