Transformer Graph Variational Autoencoder for Generative Molecular Design DOI Creative Commons

Trieu Nguyen,

Aleksandra Karolak

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Июль 23, 2024

ABSTRACT In the field of drug discovery, generation new molecules with desirable properties remains a critical challenge. Traditional methods often rely on SMILES (Simplified Molecular Input Line Entry System) representations for molecular input data, which can limit diversity and novelty generated molecules. To address this, we present Transformer Graph Variational Autoencoder (TGVAE), an innovative AI model that employs graphs as thus captures complex structural relationships within more effectively than string models. enhance capabilities, TGVAE combines Transformer, Neural Network (GNN), (VAE). Additionally, common issues like over-smoothing in training GNNs posterior collapse VAE to ensure robust improve chemically valid diverse structures. Our results demonstrate outperforms existing approaches, generating larger collection discovering structures were previously unexplored. This advancement not only brings possibilities discovery but also sets level use generation.

Язык: Английский

Artificial Intelligence (AI) Applications in Drug Discovery and Drug Delivery: Revolutionizing Personalized Medicine DOI Creative Commons
Dolores R. Serrano,

Francis C. Luciano,

Brayan J. Anaya

и другие.

Pharmaceutics, Год журнала: 2024, Номер 16(10), С. 1328 - 1328

Опубликована: Окт. 14, 2024

Artificial intelligence (AI) encompasses a broad spectrum of techniques that have been utilized by pharmaceutical companies for decades, including machine learning, deep and other advanced computational methods. These innovations unlocked unprecedented opportunities the acceleration drug discovery delivery, optimization treatment regimens, improvement patient outcomes. AI is swiftly transforming industry, revolutionizing everything from development to personalized medicine, target identification validation, selection excipients, prediction synthetic route, supply chain optimization, monitoring during continuous manufacturing processes, or predictive maintenance, among others. While integration promises enhance efficiency, reduce costs, improve both medicines health, it also raises important questions regulatory point view. In this review article, we will present comprehensive overview AI's applications in covering areas such as discovery, safety, more. By analyzing current research trends case studies, aim shed light on transformative impact industry its broader implications healthcare.

Язык: Английский

Процитировано

37
Embracing the changes and challenges with modern early drug discovery DOI
Vinay Kumar, Kunal Roy

Expert Opinion on Drug Discovery, Год журнала: 2025, Номер unknown

Опубликована: Март 17, 2025

The landscape of early drug discovery is rapidly evolving, fueled by significant advancements in artificial intelligence (AI) and machine learning (ML), which are transforming the way drugs discovered. As traditional faces growing challenges terms time, cost, efficacy, there a pressing need to integrate these emerging technologies enhance process. In this perspective, authors explore role AI ML modern discuss their application target identification, compound screening, biomarker discovery. This article based on thorough literature search using PubMed database identify relevant studies that highlight use AI/ML models computational chemistry, systems biology, data-driven approaches development. Emphasis placed how address key such as data integration, predictive performance, cost-efficiency pipeline. have potential revolutionize improving accuracy speed identifying viable candidates. However, successful integration requires overcoming related quality, model interpretability, for interdisciplinary collaboration.

Язык: Английский

Процитировано

0
Exploring 4th Generation EGFR Inhibitors: A Review of Clinical Outcomes and Structural Binding Insights. DOI

Amina Tariq,

Muhammad Shoaib,

Lingbo Qu

и другие.

European Journal of Pharmacology, Год журнала: 2025, Номер unknown, С. 177608 - 177608

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
The power of artificial intelligence for managing pandemics: A primer for public health professionals DOI Creative Commons
Martin McKee, Rikard Rosenbacke, David Stuckler

и другие.

The International Journal of Health Planning and Management, Год журнала: 2024, Номер unknown

Опубликована: Окт. 27, 2024

Abstract Artificial intelligence (AI) applications are complex and rapidly evolving, thus often poorly understood, but have potentially profound implications for public health. We offer a primer health professionals that explains some of the key concepts involved examines how these might be used in response to future pandemic. They include early outbreak detection, predictive modelling, healthcare management, risk communication, surveillance. applications, especially algorithms, ability anticipate outbreaks by integrating diverse datasets such as social media, meteorological data, mobile phone movement data. intelligence‐powered tools can also optimise delivery managing allocation resources reducing workers' exposure risks. In resource distribution, they demand logistics, while AI‐driven robots minimise physical contact settings. shows promise supporting decision‐making simulating economic impacts different policy interventions. These simulations help policymakers evaluate scenarios lockdowns allocation. Additionally, it enhance messaging, with AI‐generated communications shown more effective than human‐generated messages cases. However, there risks, privacy concerns, biases models, potential ‘false confirmations’, where AI reinforces incorrect decisions. Despite challenges, we argue will become increasingly important crises, only if integrated thoughtfully into existing systems processes.

Язык: Английский

Процитировано

3
Transformer Graph Variational Autoencoder for Generative Molecular Design DOI Creative Commons

Trieu Nguyen,

Aleksandra Karolak

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Июль 23, 2024

ABSTRACT In the field of drug discovery, generation new molecules with desirable properties remains a critical challenge. Traditional methods often rely on SMILES (Simplified Molecular Input Line Entry System) representations for molecular input data, which can limit diversity and novelty generated molecules. To address this, we present Transformer Graph Variational Autoencoder (TGVAE), an innovative AI model that employs graphs as thus captures complex structural relationships within more effectively than string models. enhance capabilities, TGVAE combines Transformer, Neural Network (GNN), (VAE). Additionally, common issues like over-smoothing in training GNNs posterior collapse VAE to ensure robust improve chemically valid diverse structures. Our results demonstrate outperforms existing approaches, generating larger collection discovering structures were previously unexplored. This advancement not only brings possibilities discovery but also sets level use generation.

Язык: Английский

Процитировано

0