Professor Charlotte Deane MBE
Professor of Structural Bioinformatics, Department of Statistics, University of Oxford
Biography: Charlotte Deane MBE is a Professor in the Department of Statistics at the University of Oxford, Executive Chair of the Engineering and Physical Sciences Research Council (EPSRC) and Co-Founder of Dalton Tx.
During the COVID-19 pandemic, she served on SAGE, the UK Government’s Scientific Advisory Group for Emergencies, and acted as UK Research and Innovation’s COVID-19 Response Director.
In 2025, Charlotte was elected as a Fellow of the International Society for Computational Biology (ISCB).
At Oxford, Charlotte leads the Oxford Protein Informatics Group (OPIG), who work on diverse problems across immunoinformatics, protein structure and small molecule drug discovery; using statistics, AI and computation to generate biological and medical insight.
Charlotte’s research focuses on the development of novel algorithms, tools and databases which are openly available to the community. They are widely used in both academia and industry and embedded in pharmaceutical drug discovery pipelines. She is a member of several advisory boards and has consulted extensively with industry, having also established a consulting arm within her research group as a way of promoting industrial interaction and use of the group’s software tools.
Charlotte is part of the team leading OpenBind, a £8 million government-backed consortium aiming to create the world’s largest open dataset of drug-protein interactions to accelerate AI-driven drug discovery. She also serves as one of five experts advising the UK Government’s new AI for Science strategy, which aims to boost AI adoption across research and accelerate scientific discovery.
Dr Srijit Seal
Visiting Scientist, Broad Institute of MIT and Harvard
Biography: Srijit Seal specializes in machine learning and cheminformatics. His research focuses on developing machine learning algorithms for drug discovery, particularly in toxicity prediction. He is also a Fellow of the Cambridge Philosophical Society and serves on the Board of Directors of the American Society for Cellular and Computational Toxicology (ASCCT). Seal received his PhD at the University of Cambridge and completed his postdoctoral training at the Broad Institute of MIT and Harvard.
Abstract: Machine learning (ML) is increasingly valuable for predicting molecular properties and toxicity in drug discovery. However, toxicity-related endpoints have always been challenging to evaluate experimentally with respect to in vivo translation due to the required resources for human and animal studies; this has impacted data availability in the field. ML can augment or even potentially replace traditional experimental processes depending on the project phase and specific goals of the prediction. For instance, models can be used to select promising compounds for on-target effects or to deselect those with undesirable characteristics (e.g., off-target or ineffective due to unfavorable pharmacokinetics). However, reliance on ML is not without risks, due to biases stemming from nonrepresentative training data, incompatible choice of algorithm to represent the underlying data, or poor model building and validation approaches. This might lead to inaccurate predictions, misinterpretation of the confidence in ML predictions, and ultimately suboptimal decision-making. Hence, understanding the predictive validity of ML models is of utmost importance to enable faster drug development timelines while improving the quality of decisions. This presentation will emphasize the need to enhance the understanding and application of machine learning models in drug discovery, focusing on well-defined data sets for toxicity prediction based on small molecule structures. We will focus on five crucial pillars for success with ML-driven molecular property and toxicity prediction: (1) data set selection, (2) structural representations, (3) model algorithm, (4) model validation, and (5) translation of predictions to decision-making. Understanding these key pillars will foster collaboration and coordination between ML researchers and toxicologists, which will help to advance drug discovery and development.
Dr Megan Houweling
Director of Science at Medstone Science- SURUS consultancy
Biography: Megan Houweling holds an MSc in Biomedical Sciences (Radboud University) and a PhD in Neuro-oncology (Amsterdam UMC) focused on combination therapies for glioblastoma. She completed the Postgraduate Education in Toxicology (PET) and is now a registered toxicologist in the Netherlands (NVT/EUROTOX). Since 2023, she has worked at Medstone Science, using AI to translate clinical and toxicological data into faster, more consistent decision-making. At Medstone Science, she applies SURUS, the company’s domain-specific NLP model that automatically categorizes and structures scientific literature. By linking SURUS with LLMs daily, this workflow delivers higher processing speed, improved traceability, and more immediate insights for risk assessment, labelling, and evidence synthesis across therapeutic areas.
Abstract: Large language models (LLMs) can turn dispersed clinical evidence into interpretable, auditable inputs for smarter drug development. This talk focuses on clinical applications (distinct from preclinical modelling) and pinpoints where LLMs add practical value today. First, LLMs enable rapid, auditable synthesis of literature and real-world evidence by retrieving, ranking, and summarizing findings across trials, labels, pharmacovigilance reports, and guidelines—supporting signal detection, benefit–risk assessment, and labelling updates. Second, retrieval-augmented generation (RAG) and tool-use pipelines allow case-level reasoning (e.g., adverse-event narratives, dechallenge/rechallenge, concomitant meds) while enforcing traceability via source citation and structured outputs. Third, LLMs can streamline protocol design and amendment-impact analyses, patient-stratification rationales, and clinician-facing explanations that improve multidisciplinary decision-making. We discuss validation strategies (prospective benchmarks, inter-rater agreement vs expert panels), safeguards (hallucination mitigation, calibration, uncertainty expression), data governance (PHI, provenance, access control), and compliance (transparency and audit documentation). Rather than replacing expert judgment, LLMs act as catalysts that compress time to insight, expand evidence coverage, and enhance consistency and explainability—ultimately integrating clinical data streams to optimize decision quality, reduce development risk, and accelerate the drug development lifecycle.
Professor Patricia B Munroe
Professor of Molecular Medicine, Centre Lead for Clinical Pharmacology and Precision Medicine, Queen Mary University of London
Biography: Patricia Munroe is a Professor of Molecular Medicine at Queen Mary University of London with a research focus on cardiovascular genetics. Her group uses genomics and computational methods to understand the molecular basis of hypertension, cardiac arrhythmias and heart failure. She has established and co-leads several international complex trait consortia, this collaborative networking has led to the discovery of 1000s of genes for disease and cardiovascular risk factors.
Ali Sarrami
Lecturer in Cardiovascular Biomechanics at the University of Manchester
Hippocrates Meets Turing – In Silico Regulatory Science in the Medical Digital Era
Dr Olivier Béquignon
Assistant Professor in AI & structure-based drug design, Leiden University
Talk title: Breaking Barriers - AI-driven cheminformatics for CNS drug discovery
Biography: Olivier J. M. Béquignon is an Assistant Professor in AI and structure-based drug discovery at Leiden University's Department of Medicinal Chemistry. He holds a PhD in Computational Drug Discovery, and a degree in Biotechnology Engineering. After a two-year postdoctoral position at Amsterdam UMC, where he studied the blood-brain barrier permeability of kinase inhibitors for glioblastoma, his research now focuses on combining AI, cheminformatics, and molecular modelling. Among his significant contributions are the 'Papyrus' bioactivity dataset, which has been downloaded more than 15,000 times, and several open-source tools for drug design.
Abstract: Predicting blood-brain barrier (BBB) permeability is a critical challenge in CNS drug discovery, often hindered by data scarcity and a lack of molecular diversity in available datasets. We present a robust machine learning framework designed to overcome these limitations and streamline the screening of brain-permeable molecules.
Our approach leverages transfer learning, pretraining models on 6 million predictions to establish a fundamental chemical understanding before fine-tuning on specific BBB tasks. We further enhance performance through multitask learning, which captures complex signal patterns more effectively than single-task approaches. Rigorous validation was conducted using a carefully designed dissimilar test set to ensure generalizability.
Results demonstrate that combining pretraining with transfer learning yields significantly higher predictive power than baseline methods. Furthermore, multitask learning consistently outperforms single-task models. This strategy offers a scalable, high-accuracy solution for prioritizing neurotherapeutics, effectively addressing the data constraints inherent in BBB research.
Dr Peter Cox
Head of the Translational Science team, Isomorphic Labs
Biography: Dr. Peter Cox is a biologist with over 25 years experience working in the pharmaceutical and biotech industries. Peter holds a PhD in Molecular Virology from the University of Glasgow, followed by postdoctoral training in Neuroscience at INSERM, Paris and the University of Cambridge.
Peter began his industry career in 1998 at Pfizer dedicating 16 years to small molecule drug discovery for chronic pain, evolving from a bench scientist to taking on significant responsibilities in pain target identification and leading drug discovery projects through to IND preparation. This period provided him with extensive experience in the molecular pharmacology of diverse target classes, innovative target identification methodologies, and comprehensive drug discovery project leadership.
In 2014, Peter joined BenevolentAI (BAI), a pioneering AI-led drug discovery company. There, he was instrumental in building and expanding a multidisciplinary team of drug discovery scientists and establishing a robust portfolio of projects. As a key member of BAI's drug discovery leadership team, he provided strategic direction, leveraged BAI's knowledge graph-based target identification platform, and led large multidisciplinary teams and drug discovery initiatives.
Peter is currently Head of the Translational Science team at Isomorphic Labs, which is developing and applying frontier AI to reimagine and advance the drug design process to unlock deeper scientific insights and faster breakthroughs. Peter’s team of in vitro pharmacologists / translational biologists are responsible for the biology strategy of Isomorphic Labs’ AI-led drug design projects and provides critical validation for their innovative in-silico predictions through the generation of experimental biological data, directly advancing the evolution of some of the most sophisticated and powerful AI drug design platforms.
