AI-driven data analysis and modelling of protein dynamics for drug discovery

This study leverages AI and computational strategies to target key membrane proteins involved in drug addiction and antimicrobial resistance, aiming to develop in silico models for drug discovery.

Our research focuses on computational strategies to target key membrane proteins involved in both drug addiction and antimicrobial resistance. This work aims to develop in silico models for drug discovery. Integrating molecular dynamics, AI, and virtual screening, this work seeks to design novel inhibitors and establish a Computational Biomedicine Unit at Brunel University of London.

Innovative solutions such as computational drug discovery offer a powerful approach to expedite the identification of new treatments for various disease models. By using AI, molecular dynamics, and virtual screening, this work aims to design inhibitors targeting key proteins involved in addiction and antimicrobial resistance (AMR), providing cost-effective, scalable solutions to these crises.

This research not only contributes to public health but also promotes ethical, sustainable drug discovery methods, reducing reliance on animal testing. By advancing therapeutic strategies for addiction and infectious diseases, this work tackles major societal challenges, offering the potential to improve quality of life and reduce economic burdens associated with these widespread issues.

Using computational methods to address drug addiction and antimicrobial resistance

This research stands out by integrating cutting-edge computational methods, such as molecular dynamics simulations, AI, and virtual screening, to address both drug addiction and antimicrobial resistance.

What sets this project apart is its dual focus on these distinct yet equally urgent health issues, using a shared computational framework to design targeted inhibitors for both. Unlike traditional drug discovery approaches, which rely heavily on experimental trials, this project leverages in silico techniques to accelerate the process, reducing costs and time

The use of AI and virtual reality in drug screening is an innovative aspect, offering a more efficient, scalable solution to complex biomedical challenges. This novel approach addresses critical gaps in understanding the structural and dynamic roles of key proteins in both addiction and AMR, pushing the boundaries of computational drug discovery.

Our approach and methods

Our research uses innovative computational strategies to target key proteins involved in both drug addiction and antimicrobial resistance (AMR). Advanced computational techniques like molecular dynamics simulations, AI, and virtual screening can be used to design novel inhibitors with the potential to revolutionise drug discovery, accelerating the development of therapies for both conditions while minimising the need for traditional experimental methods.

The dopamine transporter (DAT) plays an important role in the brain's reward system and is a key player in the development of drug addiction. Drugs of abuse, such as cocaine and methamphetamine, directly impact DAT function, leading to altered dopamine signalling and addictive behaviours. Computational methods can help to understand the structural and functional dynamics of DAT, exploring how specific small molecules or inhibitors could modulate its activity to treat addiction.

Antimicrobial Resistance (AMR) represents one of the greatest global health challenges, with Acinetobacter baumannii being a particularly dangerous pathogen due to its resistance to many antibiotics. In this area, we aim to investigate the structural and functional mechanisms of key proteins involved in the pathogen's virulence and resistance. The goal is to design computational models that will enable the identification of novel small molecules capable of disrupting these proteins' functions. By inhibiting critical proteins in A. baumannii, the research aims to reduce the pathogen's ability to cause infections and resist treatment.

Molecular dynamics will be used to study the protein-ligand interactions of DAT and proteins from A. baumannii. This method allows for the detailed analysis of how specific molecules interact with these proteins at the atomic level. High-throughput virtual screening techniques can identify potential inhibitors for their binding affinity to the target proteins. Generative AI and machine learning algorithms will be applied to predict and optimise drug properties, enhancing the efficiency of the drug discovery process.

This research addresses critical public health issues. These conditions have significant societal and economic impacts, and the development of new therapeutic strategies could save lives, reduce healthcare costs, and improve quality of life. By focusing on computational methods, the project reduces reliance on animal testing, supporting the use of ethical, sustainable approaches in biomedical research.

Publications

• M.A. Sahai, J. Opacka-Juffry. Molecular mechanisms of action of stimulant novel psychoactive substances that target the high-affinity transporter for dopamine, Neuronal Signal. 5 (2021). doi: 10.1042/NS20210006
• B. Loi, M.A. Sahai, M.A. De Luca, H. Shiref, J. Opacka-Juffry, The Role of Dopamine in the Stimulant Characteristics of Novel Psychoactive Substances (NPS)—Neurobiological and Computational Assessment Using the Case of Desoxypipradrol (2-DPMP), Front. Pharmacol. 11 (2020) 806. doi:10.3389/fphar.2020.00806
• M.A. Sahai, C. Davidson, N. Dutta, J. Opacka-Juffry, Mechanistic Insights into the Stimulant Properties of Novel Psychoactive Substances (NPS) and Their Discrimination by the Dopamine Transporter—In Silico and In Vitro Exploration of Dissociative Diarylethylamines, Brain Sci. 8 (2018) 63. doi:10.3390/brainsci8040063
• M.A. Sahai, C. Davidson, G. Khelashvili, V. Barrese, N. Dutta, H. Weinstein, J. Opacka-Juffry, Combined in vitro and in silico approaches to the assessment of stimulant properties of novel psychoactive substances – The case of the benzofuran 5-MAPB, Prog. Neuro-Psychopharmacology Biol. Psychiatry. 75 (2017) 1–9. doi:10.1016/j.pnpbp.2016.11.004
• S Motta, L. Callea, L. Bonati, A. Pandini, PathDetect-SOM: A Neural Network Approach for the Identification of Pathways in Ligand Binding Simulations. J Chem Theory Comput 18 (2022), 1957–1968. doi:10.1021/acs.jctc.1c01163