T-ALPHA: A Hierarchical Transformer-Based Deep Neural Network for Protein-Ligand Binding Affinity Prediction With Uncertainty-Aware Self-Learning for Protein-Specific Alignment DOI Creative Commons
Gregory W. Kyro, Anthony M. Smaldone, Yu Shee

et al.

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

Published: Dec. 20, 2024

Abstract There is significant interest in targeting disease-causing proteins with small molecule inhibitors to restore healthy cellular states. The ability accurately predict the binding affinity of molecules a protein target silico enables rapid identification candidate and facilitates optimization on-target potency. In this work, we present T-ALPHA, novel deep learning model that enhances protein-ligand prediction by integrating multimodal feature representations within hierarchical transformer framework capture information critical predicting affinity. T-ALPHA outperforms all existing models reported literature on multiple benchmarks designed evaluate scoring functions. Remarkably, maintains state-of-the-art performance when utilizing predicted structures rather than crystal structures, powerful capability real-world drug discovery applications where experimentally determined are often unavailable or incomplete. Additionally, an uncertainty-aware self-learning method for protein-specific alignment does not require additional experimental data, demonstrate it improves T-ALPHA’s rank compounds biologically targets such as SARS-CoV-2 main protease epidermal growth factor receptor. To facilitate implementation reproducibility results presented paper, have made our software available at https://github.com/gregory-kyro/T-ALPHA .

Language: Английский

A Review of Large Language Models and Autonomous Agents in Chemistry DOI Creative Commons
Mayk Caldas Ramos, Christopher J. Collison, Andrew Dickson White

et al.

Chemical Science, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 9, 2024

Large language models (LLMs) have emerged as powerful tools in chemistry, significantly impacting molecule design, property prediction, and synthesis optimization. This review highlights LLM capabilities these domains their potential to accelerate scientific discovery through automation. We also LLM-based autonomous agents: LLMs with a broader set of interact surrounding environment. These agents perform diverse tasks such paper scraping, interfacing automated laboratories, planning. As are an emerging topic, we extend the scope our beyond chemistry discuss across any domains. covers recent history, current capabilities, design agents, addressing specific challenges, opportunities, future directions chemistry. Key challenges include data quality integration, model interpretability, need for standard benchmarks, while point towards more sophisticated multi-modal enhanced collaboration between experimental methods. Due quick pace this field, repository has been built keep track latest studies: https://github.com/ur-whitelab/LLMs-in-science.

Language: Английский

Citations

13
T-ALPHA: A Hierarchical Transformer-Based Deep Neural Network for Protein–Ligand Binding Affinity Prediction with Uncertainty-Aware Self-Learning for Protein-Specific Alignment DOI
Gregory W. Kyro, Anthony M. Smaldone, Yu Shee

et al.

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

Published: Feb. 18, 2025

There is significant interest in targeting disease-causing proteins with small molecule inhibitors to restore healthy cellular states. The ability accurately predict the binding affinity of molecules a protein target silico enables rapid identification candidate and facilitates optimization on-target potency. In this work, we present T-ALPHA, novel deep learning model that enhances protein–ligand prediction by integrating multimodal feature representations within hierarchical transformer framework capture information critical predicting affinity. T-ALPHA outperforms all existing models reported literature on multiple benchmarks designed evaluate scoring functions. Remarkably, maintains state-of-the-art performance when utilizing predicted structures rather than crystal structures, powerful capability real-world drug discovery applications where experimentally determined are often unavailable or incomplete. Additionally, an uncertainty-aware self-learning method for protein-specific alignment does not require additional experimental data demonstrate it improves T-ALPHA's rank compounds biologically targets such as SARS-CoV-2 main protease epidermal growth factor receptor. To facilitate implementation reproducibility results presented paper, made our software available at https://github.com/gregory-kyro/T-ALPHA.

Language: Английский

Citations

1
Effective drug-target affinity prediction via generative active learning DOI
Yuansheng Liu, Zhenran Zhou, Xiaofeng Cao

et al.

Information Sciences, Journal Year: 2024, Volume and Issue: 679, P. 121135 - 121135

Published: July 3, 2024

Language: Английский

Citations

7
Activity cliff-aware reinforcement learning for de novo drug design DOI Creative Commons
Xiuyuan Hu, Guoqing Liu, Yang Zhao

et al.

Journal of Cheminformatics, Journal Year: 2025, Volume and Issue: 17(1)

Published: April 21, 2025

Language: Английский

Citations

0
Molecular property prediction using pretrained-BERT and Bayesian active learning: a data-efficient approach to drug design DOI Creative Commons
Muhammad Arslan Masood, Samuel Kaski, Tianyu Cui

et al.

Journal of Cheminformatics, Journal Year: 2025, Volume and Issue: 17(1)

Published: April 23, 2025

Language: Английский

Citations

0
CardioGenAI: A Machine Learning-Based Framework for Re-Engineering Drugs for Reduced hERG Liability DOI
Gregory W. Kyro, Matthew T. Martin, Eric D. Watt

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 10, 2024

Abstract The link between in vitro hERG ion channel inhibition and subsequent vivo QT interval prolongation, a critical risk factor for the development of arrythmias such as Torsade de Pointes, is so well established that activity alone often sufficient to end an otherwise promising drug candidate. It therefore tremendous interest develop advanced methods identifying hERG-active compounds early stages development, proposing redesigned with reduced liability preserved primary pharmacology. In this work, we present CardioGenAI, machine learning-based framework re-engineering both developmental commercially available drugs while preserving their pharmacological activity. incorporates novel state-of-the-art discriminative models predicting activity, against voltage-gated NaV1.5 CaV1.2 channels due potential implications modulating arrhythmogenic induced by blockade. We applied complete pimozide, FDA-approved antipsychotic agent demonstrates high affinity channel, generated 100 refined candidates. Remarkably, among candidates fluspirilene, compound which same class pimozide (diphenylmethanes) has similar yet exhibits over 700-fold weaker binding hERG. Furthermore, demonstrated framework's ability optimize hERG, profiles multiple maintaining physicochemical nature original drugs. envision method can effectively be exhibiting liabilities provide means rescuing programs have stalled hERG-related safety concerns. Additionally, also serve independently effective components virtual screening pipelines. made all our software open-source at https://github.com/gregory-kyro/CardioGenAI facilitate integration CardioGenAI molecular hypothesis generation into discovery workflows.

Language: Английский

Citations

3
CardioGenAI: a machine learning-based framework for re-engineering drugs for reduced hERG liability DOI Creative Commons
Gregory W. Kyro, Matthew T. Martin, Eric D. Watt

et al.

Journal of Cheminformatics, Journal Year: 2025, Volume and Issue: 17(1)

Published: March 5, 2025

The link between in vitro hERG ion channel inhibition and subsequent vivo QT interval prolongation, a critical risk factor for the development of arrythmias such as Torsade de Pointes, is so well established that activity alone often sufficient to end an otherwise promising drug candidate. It therefore tremendous interest develop advanced methods identifying hERG-active compounds early stages development, proposing redesigned with reduced liability preserved primary pharmacology. In this work, we present CardioGenAI, machine learning-based framework re-engineering both developmental commercially available drugs while preserving their pharmacological activity. incorporates novel state-of-the-art discriminative models predicting activity, against voltage-gated NaV1.5 CaV1.2 channels due potential implications modulating arrhythmogenic induced by blockade. We applied complete pimozide, FDA-approved antipsychotic agent demonstrates high affinity channel, generated 100 refined candidates. Remarkably, among candidates fluspirilene, compound which same class pimozide (diphenylmethanes) has similar yet exhibits over 700-fold weaker binding hERG. Furthermore, demonstrated framework's ability optimize hERG, profiles multiple maintaining physicochemical nature original drugs. envision method can effectively be exhibiting liabilities provide means rescuing programs have stalled hERG-related safety concerns. Additionally, also serve independently effective components virtual screening pipelines. made all our software open-source at https://github.com/gregory-kyro/CardioGenAI facilitate integration CardioGenAI molecular hypothesis generation into discovery workflows.Scientific contributionThis work introduces designed re-engineer facing challenges. addition, function

Language: Английский

Citations

0
Generative AI in drug discovery and development: the next revolution of drug discovery and development would be directed by generative AI DOI Open Access
Chiranjib Chakraborty, Manojit Bhattacharya, Soumen Pal

et al.

Annals of Medicine and Surgery, Journal Year: 2024, Volume and Issue: 86(10), P. 6340 - 6343

Published: Aug. 14, 2024

Language: Английский

Citations

2
T-ALPHA: A Hierarchical Transformer-Based Deep Neural Network for Protein-Ligand Binding Affinity Prediction With Uncertainty-Aware Self-Learning for Protein-Specific Alignment DOI Creative Commons
Gregory W. Kyro, Anthony M. Smaldone, Yu Shee

et al.

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

Published: Dec. 20, 2024

Abstract There is significant interest in targeting disease-causing proteins with small molecule inhibitors to restore healthy cellular states. The ability accurately predict the binding affinity of molecules a protein target silico enables rapid identification candidate and facilitates optimization on-target potency. In this work, we present T-ALPHA, novel deep learning model that enhances protein-ligand prediction by integrating multimodal feature representations within hierarchical transformer framework capture information critical predicting affinity. T-ALPHA outperforms all existing models reported literature on multiple benchmarks designed evaluate scoring functions. Remarkably, maintains state-of-the-art performance when utilizing predicted structures rather than crystal structures, powerful capability real-world drug discovery applications where experimentally determined are often unavailable or incomplete. Additionally, an uncertainty-aware self-learning method for protein-specific alignment does not require additional experimental data, demonstrate it improves T-ALPHA’s rank compounds biologically targets such as SARS-CoV-2 main protease epidermal growth factor receptor. To facilitate implementation reproducibility results presented paper, have made our software available at https://github.com/gregory-kyro/T-ALPHA .

Language: Английский

Citations

0