Sequential biomedical data is ubiquitous, from time-resolved data about patient encounters in the clinical realm to DNA sequences in the biological domain.  The talk will review our latest work in representation learning from longitudinal data, with a particular focus on finding optimal representations for complex and sparse healthcare data. We show how these representations are useful for comparing patient journeys and finding patients with similar health outcomes. We will also venture into the field of genome engineering, where we build models that work on DNA sequences for predicting editing outcomes for base and prime editors. 

Accurate interpretation of medical images is crucial for disease diagnosis and treatment, and AI has the potential to minimize errors, reduce delays, and improve accessibility. The focal point of this presentation lies in a grand ambition: the development of 'Generalist Medical AI' systems that can closely resemble doctors in their ability to reason through a wide range of medical tasks, incorporate multiple data modalities, and communicate in natural language. Starting with pioneering algorithms that have already demonstrated their potential in diagnosing diseases from chest X-rays or electrocardiograms, matching the proficiency of expert radiologists and cardiologists, I will delve into the core challenges and advancements in the field. The discussion will navigate towards the topic of label-efficient AI models: with a scarcity of meticulously annotated data in healthcare, the development of AI systems capable of learning effectively from limited labels has become a key concern. In this vein, I'll delve into how the innovative use of self-supervision and pre-training methods has led to algorithmic advancements that can perform high-level diagnostic tasks using significantly less annotated data. Additionally, I will talk about initiatives in data curation, human-AI collaboration, and the creation of open benchmarks to evaluate the generalizability of medical AI algorithms. In sum, this talk aims to deliver a comprehensive picture of the state of 'Generalist Medical AI,' the advancements made, the challenges faced, and the prospects lying ahead.

Chronic hepatitis B virus (HBV) infection leads to liver damage that increases the risk of hepatocellular carcinoma and liver cirrhosis. Individuals with chronic HBV infection are often either treated with interferon alpha or nucleoside reverse transcriptase inhibitors (NTRL). While these NTRLs inhibit de novo DNA synthesis, they do not represent a functional cure for chronic HBV infection and so must be taken indefinitely. The resulting life-long treatment leads to an increased risk of selection for treatment resistant strains of HBV. Consequently, there is increased interest in a novel treatment modality, capsid protein allosteric modulators (CPAMs), that blocks a crucial step in the viral life cycle. I'll discuss recent work that identifies HBV serum RNA as an informative biomarker of CPAM treatment efficacy, evaluates CPAMs as a potential functional cure for HBV infection, and illustrates the role of mechanistic modelling in trial design using an age structured, multi-scale mathematical model. 

A major reason for the failure of cancer treatment and disease progression is the heterogeneous composition of tumor cells at the genetic, epigenetic, and phenotypic levels. Despite extensive efforts to characterize the makeup of individual cells, there is still much to be learned about the interactions between heterogeneous cancer cells and between cancer cells and the microenvironment in the context of cancer invasion. Clinical studies and in vivo models have shown that cancer invasion predominantly occurs through collective invasion packs, which invade more aggressively and result in worse outcomes. In vitro experiments on non-small cell lung cancer spheroids have demonstrated that the invasion packs consist of leaders and followers who engage in mutualistic social interactions during collective invasion. Many fundamental questions remain unanswered: What is the division of labor within the heterogeneous invasion pack? How does the leader phenotype emerge? Are the phenotypes plastic? What's the role of the individual "cheaters"? How does the invasion pack interact with the stroma? Can the social interaction network be exploited to devise novel treatment strategies? I will discuss recent modeling efforts to address these questions and hope to convince you that identifying and perturbing the "weak links" within the social interaction network can disrupt collective invasion and potentially prevent the malignant progression of cancer. 

Mathematical models are useful tools for guiding infectious disease outbreak control measures. Before a pathogen has even entered a host population, models can be used to determine the locations that are most at risk of outbreaks, allowing limited surveillance resources to be deployed effectively. Early in an outbreak, key questions for policy advisors include whether initial cases will lead on to a major epidemic or fade out as a minor outbreak. When a major epidemic is ongoing, models can be applied to track pathogen transmissibility and inform interventions. And towards the end of (or after) an outbreak, models can be used to estimate the probability that the outbreak is over and that no cases will be detected in future, with implications for when interventions can be lifted safely. In this talk, I will summarise the work done by my research group on modelling different stages of infectious disease outbreaks. This includes: i) Before an outbreak: Projections of the locations at-risk from vector-borne pathogens towards the end of the 21^st century under a changing climate; ii) Early in an outbreak: Methods for estimating the risk that introduced cases will lead to a major epidemic; and iii) During a major epidemic: A novel approach for inferring the time-dependent reproduction number during outbreaks when disease incidence time series are aggregated temporally (e.g. weekly case numbers are reported rather than daily case numbers). In addition to discussing this work, I will suggest areas for further research that will allow effective interventions to be planned during future infectious disease outbreaks.

Modern medicine has given us effective tools to treat some of the most significant and burdensome diseases. At the same time, it is becoming consistently more challenging and more expensive to develop new therapeutics. A key factor in this trend is that we simply don't understand the underlying biology of disease, and which interventions might meaningfully modulate clinical outcomes and in which patients. To achieve this goal, we are bringing together large amounts of high content data, taken both from humans and from human-derived cellular systems generated in our own lab. Those are then used to learn a meaningful representation of biological states via cutting edge machine learning methods, which enable us to make predictions about novel targets, coherent patient segments, and the clinical effect of molecules. Our ultimate goal is to develop a new approach to drug development that uses high-quality data and ML models to design novel, safe, and effective therapies that help more people, faster, and at a lower cost. 

Mathematical and computational techniques can improve our understanding of diseases. In this talk, I’ll present ways in which data from cancer patients can be combined with mathematical modelling and used to improve cancer treatments.

Given the variability in individual responses to cancer treatments, agent-based modelling has been a useful technique for accurately capturing cellular behaviours that may lead to stochasticity in patient outcomes. Using a hybrid agent-based model and partial differential equation system, we developed a model for brain cancer (glioblastoma) growth informed by ex-vivo patient samples. Extending the model to capture patient treatment with an oncolytic virus rQNestin, we used our model to propose reasons for treatment failure, which was later confirmed with further patient samples. More recently, we extended this model to investigate the effectiveness of combination treatments (chemotherapy, virotherapy and immunotherapy) informed by individual patient imaging mass cytometry.

This talk hopes to provide examples of ways mathematical and computational modelling can be used to run “virtual” clinical trials with the goal of obtaining more effective treatments for diseases.  

Many natural and social phenomena involve individual agents coming together to create group dynamics, whether the agents are drivers in a traffic jam, cells in a developing tissue, or locusts in a swarm. Here I will focus on two examples of such emergent behavior in biology, specifically cell interactions during pattern formation in zebrafish skin and gametophyte development in ferns. Different modeling approaches provide complementary insights into these systems and face different challenges. For example, vertex-based models describe cell shape, while more efficient agent-based models treat cells as particles. Continuum models, which track the evolution of cell densities, are more amenable to analysis, but it is often difficult to relate their few parameters to specific cell interactions. In this talk, I will overview our models of cell behavior in biological patterns and discuss our ongoing work on quantitatively relating different types of models using topological data analysis and data-driven techniques.