QMUL Centre for Probability, Statistics and Data Science News

Why averages fail for bacteria in the open ocean

Tue, 10 Mar 2026 00:00:00 +0100

How can bacteria that forage on organic particles survive in vast ocean regions where such particles are extremely sparse? A new study by researchers from Queen Mary University of London and ETH Zurich shows that variability at the level of individual bacteria plays a central role. Using a probabilistic population model linking mathematics and microbiology, the team demonstrates that rare, high-impact encounters sustain bacterial populations even when average conditions suggest decline. Sinking organic particles, often called "marine snow", transport carbon from the surface to the deep ocean, accounting for roughly half of the ocean's total organic carbon export. As these particles sink, bacteria colonise and degrade them, influencing how much carbon ultimately reaches the deep ocean. For bacteria that rely on these organic particles, the open ocean presents a major challenge. After leaving one particle, they may spend long periods searching for the next one. Reasoning based on average encounter times has long suggested that this "particle hopping" strategy should rarely succeed, as most individual bacteria would exhaust their energy before reaching another particle. When averages fall short To probe this paradox more deeply, the research team combined environmental microbiology with stochastic modelling to derive a generalized branching-process model of bacterial foraging. Rather than relying on average encounter rates, the model represents the distribution of search times and particle interaction outcomes, including rare rapid encounters with large particles that generate many offspring. Although such encounters are infrequent, their outsized contribution can reverse what averages would predict for population growth. The authors describe this effect as "stochastic resilience". The results indicate that variability allows motile, particle-foraging bacteria to persist across a much broader range of marine environments than average-based reasoning would suggest, including parts of the open ocean and deep (bathypelagic) waters. Crucially, this persistence does not require bacteria to endure years-long periods of starvation. While many individuals fail to find another particle, a small number succeed quickly and generate enough offspring to sustain the population. Variability that shapes ocean-scale processes It is still difficult for ocean scientists to connect tiny microbial behaviours, such as movement or attachment to particles, to the much larger ecological and chemical processes in the ocean. Large-scale models necessarily describe microbial dynamics in terms of aggregate or average rates. Yet variability and rare events can influence population-level outcomes in ways that mean behaviour alone does not capture. By combining mathematical theory with environmental microbiology, this study shows how probabilistic frameworks can make those effects explicit. Sinking particles form the backbone of the ocean's biological carbon pump, transporting organic matter toward the deep sea. As bacteria colonize and degrade these particles, they influence how much carbon is respired back to CO₂ and how much continues downward. By highlighting encounter variability, the study suggests that models based solely on average rates may underestimate microbial contributions to carbon export. Dr Natasha Blitvic, Reader in Mathematical Sciences at Queen Mary, said: "It's counterintuitive that populations can persist when most individual searches fail. Rare rapid encounters make the difference, but you only see that when you step away from averages and think probabilistically. Bringing mathematics and marine microbiology together made that perspective possible." Professor Roman Stocker, Head of the Environmental Microfluidics Lab at ETH Zurich, added: "We have investigated the movement behaviour of marine bacteria quite extensively to date, but it was only thanks to the mathematical framework that Natasha and Vicente contributed that we were together finally able to scale up the effect of that behaviour from single cells to the ocean ecosystem and thereby demonstrate that the probabilistic nature of that behaviour is a key ingredient in understanding microbial foraging strategies and ultimately the important ecological and biogeochemical processes they underpin". Stochastic resilience enables particle foraging in oligotrophic marine environments, published on PNAS can be viewed here: www.pnas.org/doi/10.1073/pnas.2508238123

Three women leading research in Science and Engineering

Sun, 01 Mar 2026 00:00:00 +0100

To mark International Women's Day, we spotlight three women leading research in Science and Engineering. Meet Silvia Liverani (Head of the Centre for Probability, Statistics and Data Science) Mona Jaber (Head of the Centre for Networks, Communications, and Systems) and Ana Sobrido (Head of the Centre for Centre for Sustainable Engineering). Silvia Liverani I love uncovering the structure hidden in messy data using statistical models to find those patterns. My research focuses on developing advanced statistical methods for complex datasets, and one aspect I really enjoy is that I get to collaborate with researchers in other fields, including biologists, psychologists, clinicians, etc. I have been the Head of the Centre in Probability, Statistics and Data Science since 2023. The Centre is an exciting group of academics, PDRAs and PhD students, spanning from pure mathematics to applied statistics and image processing. Mona Jaber As a child, I was fascinated by how engineering could solve real-world problems and I knew then that I wanted to be an engineer. That early curiosity has grown into a lifelong passion that now drives my research in artificial intelligence and Internet of Things technologies, accelerating our progress toward more sustainable urban environments. Today, as Head of the Centre for Networks, Communications, and Systems, I have the privilege of working alongside exceptional colleagues on truly inspiring projects. What excites me most is not only the individual breakthroughs, but the possibility of bringing these innovations together, connecting ideas, systems, and disciplines, to then help shape a more sustainable future. Ana Sobrido My love for science started as a kid in school, when watching chemical reactions unfold felt like witnessing magic, sparking a curiosity to understand the hidden rules behind them. This led me to pursue a career in Chemistry where I developed a particular interest in materials for energy. My research pioneers sustainable materials and innovative manufacturing approaches to enable the next generation of energy storage and conversion technologies, helping accelerate the transition to a more resilient and low-carbon future. As Head of the Centre for Sustainable Engineering, I am excited about building on strong existing foundations to address the most pressing sustainability challenges. I look forward to working with fantastic colleagues, driving innovation and impactful research and excellence in teaching and learning, while fostering an open, inclusive, and inspiring environment for all.

Probabilistic modelling enables an experimental method in a study of how bacteria exchange DNA, ...

Tue, 09 Dec 2025 00:00:00 +0100

As part of a team of collaborators, Dr. Natasha Blitvic recently published an article at the Proceedings of the National Academy of Sciences of the United States of America entitled "Fluid flow generates bacterial conjugation hot spots by increasing the rate of shear-driven cell–cell encounters". The study uses probabilistic modelling and analysis to an experimental study of how shear flow conditions affect the rate at which bacteria exchange DNA. This research was supported by a 2024 Simons Foundation Pivot Fellowship - Dr Blitvic is the first mathematician to win the accolade - which allowed her to spend a year embedded in microbial ecology et ETH Zurich. She presented these results at the Pivot Annual Meeting. A pure mathematician by training, Dr. Blitvic is also interested in ways in which probabilistic reasoning can provide key insights in other disciplines. She looks for unexpected probabilistic intuition in areas such as combinatorics and algebra, and brings those to bear on hard open problems. Apart from the aforementioned study, during the fellowship year, Dr. Blitvic collaborated on a number of projects, such as on ongoing study of how rare events provide resilience to marine microbial populations in nutrient-deplete environments, more on this to follow soon.

Two-day workshop in "Geometric Methods in Probability"

Wed, 03 Dec 2025 00:00:00 +0100

Our centre members Dr. Amaranta Membrillo Solis and Prof. John Moriarty are co-organising a two-day workshop in the School of Mathematical Sciences on the 12th and 13th January, titled "Geometric Methods in Probability". The first event on Monday 12th January will consist of a London Probability Day (LPD), an one-day conference in probability. The second event Tuesday 13th January will consist of a research workshop as part of the Tangents and New Normals (TnN) network. The latter is a new international research network led by Dr. Membrillo Solis and Prof. Moriarty, set to explore how geometry can unlock new frontiers in probability, statistics, and machine learning. Please see the workshop web page for more details.

Dr. Nicolás Hernández awarded LMS Grant to establish a UK Functional Data Analysis group

Wed, 03 Dec 2025 00:00:00 +0100

Dr. Nicolás Hernández, Lecturer in Statistics at our Centre, has been awarded a Scheme 3 grant by the London Mathematical Society. The grant will support the creation of a new Joint Research Group focused on "Statistical Modelling and Inference for Functional Data Analysis". Functional Data Analysis is a rapidly growing area in statistics, yet the UK community currently lacks a dedicated, formal space for regular interaction. This new Joint Research Group aims to address this by bringing together theoretical and applied statisticians from across the country to build a stronger, more cohesive research network. The project is a collaboration between Dr. Hernández and Dr. Haiyan Liu (University of Leeds) and Prof. Karthik Bharath (University of Nottingham). The group will organize three meetings in the upcoming academic year, hosted at QMUL, the University of Leeds, and the University of Nottingham.

Meet the PhD graduates of our Centre

Wed, 26 Nov 2025 00:00:00 +0100

Our Centre currently hosts a vibrant community of young researchers with 23 PhD students working under the supervision of the Centre's academics. But what career paths do our PhD students follow after graduation? In this page, you can find information about our recent PhD graduates, including the topic of their theses, and their subsequent role.

A recent study by Eleni Matechou receives an award by the Biometrics Society

Wed, 22 Oct 2025 23:00:00 +0100

Dr Ioannis Rotous announced winner of 2025 Young Biometrician Award for a recent paper "A Pólya Tree modelling framework for batch-mark data", co-authored by Dr Eleni Matechou and published in the Annals of Applied Statistics. The judging committee wrote: "This paper presents a novel Bayesian nonparametric framework for analysing batch-mark data, where animals are marked in groups rather than uniquely, using a Pólya Tree prior on a latent entry–exit grid. The method enables exact and computationally efficient inference on population size and survival in open populations, scaling with the number of sampling occasions rather than individuals. Through simulation studies and two real-world ecological applications, the approach demonstrates clear advantages over existing approximate or computationally intensive models, providing a flexible and general framework for ecological monitoring. The panel selected this paper for its methodological innovation and clear practical impact. Ioannis Rotous's work offers a major advance in modelling complex ecological data, combining statistical rigour with real-world applicability. The paper stands out for its originality, technical depth, and strong applied relevance—qualities that make it a deserving recipient of the Young Biometrician Award." Dr Matechou, who supervised Dr Rotous for his PhD, stated: "It's been a real pleasure to see Ioannis's work evolve from early PhD ideas into a published paper and now an award-winning contribution to statistics." For the official announcement, see the Biometric Society post.

From biodiversity to artificial intelligence

Wed, 15 Oct 2025 23:00:00 +0100

We showcase the research work of Kabiru Abubakari, a PhD student in the Centre for Probability, Statistics and Data Science, and the research work of Dr David Mguni, a lecturer in the Centre for Multimodal AI. Kabiru Abubakari Kabiru's research focuses on Bayesian spatial modelling for biodiversity. His PhD project is devoted to developing and applying Bayesian spatial and spatio-temporal modelling techniques to enhance understanding of the association between plant species at risk of extinction and areas in need of protection in the face of climate change, changing land use (especially agriculture), and pollution. Working together with his supervisors — Prof Silvia Liverani (SMS), Prof Andrew Leitch (SBBS), and Dr Ilia Leitch (Royal Botanic Garden, Kew) — Kabiru combines statistical modelling and ecology to develop methods that better capture uncertainty in biodiversity data. His academic journey began with a degree in Economics at the University for Development Studies (UDS) in Tamale, Ghana, where he graduated in 2020. Since joining Queen Mary, Kabiru has also been very active in supporting students of Black heritage as a tutor in Levelling Up Maths and as a panellist at the Black Heroes of Mathematics Conference. Read more about Kabiru's research in this poster. David Mguni David is a Lecturer in Artificial Intelligence. His research spans reinforcement learning, game theory, and optimal control, with a focus on developing self-improving, cooperative learning systems. His work contributes to a broader vision of building AI that can reason, adapt, and learn autonomously in an open-ended world. Together with his PhD student Yaqi Sun and master's students, David is working towards one of the grand goals of artificial intelligence: creating systems that can not only learn from existing training data but also learn how to learn and invent their own challenges. The group's research on the Recursive Meta-Learning Framework explores how intelligent systems can evolve their own learning rules and generate and solve new problems that push them beyond the limits of human-derived data. A central focus of the group's work is reinforcement learning — particularly understanding how multiple intelligent systems can cooperate, compete, and coordinate in open, dynamic environments. The group's research seeks to overcome the limitations of traditional reinforcement learning algorithms by enabling AI to learn the rules of learning itself. This approach has far-reaching implications. By allowing AI systems to invent new challenges, discover hidden structures, and maintain stability as they learn together, the research moves toward the long-term goal of artificial general intelligence: machines capable of generalising knowledge, adapting creatively, and cooperating safely across domains. Possible applications range from AI programs that autonomously generate novel mathematical proofs to agents that continually refine their understanding of molecular structures for drug discovery. The group's work blends theory with practical experimentation, drawing on dynamical systems, game theory, category theory, stochastic control, and variational optimisation. These mathematical foundations ensure that the learning mechanisms they develop are not only powerful and flexible but also grounded in principles that make them interpretable, stable, and safe. To learn more about David's research, read also here.

Developing mathematical foundations for understanding the dynamics of multiscale systems

Sun, 05 Oct 2025 23:00:00 +0100

Dr Amaranta Mebrillo Solis recently attended the kick-off meeting of the COST Action "multiscale Stochastics, Patterns, and Analysis of Combinatorial Environments" (mSPACE) in Brussels. COST (European Cooperation in Science and Technology) is a funding organisation for research and innovation networks. The Actions (funded projects) help connect research initiatives across Europe and beyond and enable researchers and innovators to grow their ideas in science and technology. More than 40 researchers in 21 countries across Europe are currently members of mSPACE, and Dr Membrillo Solis will be a co-leader of one of its four working groups. The Action is funded for four years and is expected to attract researchers worldwide interested in developing a rigorous mathematical foundation for understanding the dynamics of multiscale systems. Dr Membrillo Solis shared her impression from the kick-off meeting: "I felt excited to be part of this project. It is the first time I have been involved in such a large international network for research. Also, it was inspiring to learn from the diverse expertise of participants, ranging from early-career researchers to senior and established mathematicians, all of whom are eager to collaborate. I also had the opportunity to connect this project with a UK network in metric geometry and analysis, creating synergies between two groups whose research interests and goals align. From the network, I have also secured funds to co-organise a workshop in the UK this May, focused on spectral geometry and optimal transport. I am confident that these initiatives will strengthen both my work at QMUL and its visibility on international platforms."

Engaging undergraduate students with research via the STRIDE programme: An example from our Centre

Mon, 22 Sep 2025 23:00:00 +0100

The Summer Training Research Initiative to Support Diversity and Equality (STRIDE) programme is an initiative of Queen Mary aiming to provide opportunities to students from underrepresented groups to engage with a research project. The programme lasts for 8 weeks in the summer months during which the students are also provided with financial support. This summer the Centre for Probability, Statistics and Data Science hosted Catalina Neascu, a third year undergraduate student who worked under the supervision of Dr. Nicolás Hernández. The project focused on Bayesian Optimisation to identify the most important time segment within an ECG signal for diagnosing heart conditions like ischemia. By focusing on this key time interval, the method aims to improve diagnostic accuracy and boost classification algorithms. Catalina was excited to share her experience and commented: "I decided to apply to the summer internship for two main reasons. One is that I really enjoyed working on my undergraduate project and the internship would offer me another opportunity to experience research, and not only, to delve deeper into the research I started with that project. I have also decided to go into a PhD path, and the summer internship offered me an opportunity to see how it is to work on a more "loose" project, compared to my undergraduate dissertation, that could go into multiple directions and presented me with decisions to be made. The internship was challenging in the sense of practical skills, having to apply new techniques and ideas which I have not even come across during my time as undergraduate. I did greatly enjoy the process of trying to implement different theories, working on solving issues that might come up, I even enjoyed when my Python code would crash! Dr. Hernandez made the entire process easier, helping me through any kinks and problems that I could not solve by myself. I really enjoyed my time as a research intern, and it has solidified my inclination for getting a PhD."

A busy conference summer for the academics of the Centre

Mon, 15 Sep 2025 23:00:00 +0100

Summer is typically a period where many conferences take place keeping academics busy. From organising events to giving talks to conferences, members of the Centre were active during the 2025 summer, presenting their research and networking with the UK and the international community. Natasha Blitvic was an invited plenary speaker at the 15th Nordic Combinatorial Conference that took place at Reykjavik, the picturesque capital of Iceland in June. Additionally Natasha was also a plenary speaker at the International Conference on Permutation Patterns 2025, St Andrews, Scotland, in July. Primoz Skraba and Omer Bobrowski coorganised the Erlangen AI Hub Conference 2025, which took place here at the School of Mathematical Sciences, QMUL. The event was under the auspices of the UKRI AI Hub "Mathematical Foundations of Intelligence: An Erlangen Programme for AI " of which Primoz and Omer are principal investigators. Omer was additionally an invited speaker at a series of conferences: Dynamics, Topology and Computations in Będlewo, Poland, in June, The Geometric Realization of AATRN (Applied Algebraic Topology Research Network), in Chicago, USA, in August, and Workshop on Random Simplicial Complexes and Hypergraphs, Nottingham-Trent University, in September. He also gave a mini-course titled "Random topology – theory and applications" at the workshop Topological data analysis in stochastic geometry and image processing, Aarhus, Denmark, in August. Sasha Gnedin was an invited speaker in two conferences at the Fields Institute, Toronto, Canada in May: Fields-CFI Conference on Optimal Stopping and Its Applications in Finance and Insurance, and International Conference on Probabilistic, Combinatorial and Asymptotic Methods for the Analysis of Algorithms (AofA 2025). He also gave a talk at the 22nd International Conference on Random Structures & Algorithms, in Vienna, in August. Kostas Papafitsoros was coorganiser of the 10th International Conference on Scale Space and Variational Methods in Computer Vision (SSVM 2025), which took place in Dartington Hall, Devon, in May. Kostas also participated at the European Conference on Numerical Mathematics and Advanced Applications, ENUMATH 2025, where he coorganised a session on "Learning PDE-based Models from Data" and also giving a talk there. Alex Shestopaloff was coorganiser of The Conference on Networks and Graphs which took place at Belgrade, Serbia, in August. Our recent graduate Adam Onus, who will soon be leaving QMUL for a postfoc at Oxford, gave a talk at the 11th Conference in the Algebraic Topology: Methods, Computation & Science, Montana, USA, in July and presented a poster at the joint British Mathematical Colloquium and British Applied Mathematics Colloquium, Exeter, in June, which won the PhD Student Poster Prize! Finally, Eftychia Solea and Adam Onus were coorganisers of the PiFORUM25 conference which took place at QMUL in September. This is an annual conference run by the Piscopia Initiative, which is aimed at encouraging women and gender minorities to pursue PhD's and/or careers in mathematics.

Using mathematics to study microbe growth, movements and interactions

Wed, 10 Sep 2025 23:00:00 +0100

As a recipient of a 2024 Simons Foundation Pivot Fellowship, Dr. Natasha Blitvic recently participated at the Foundation's spring gathering of Pivot Fellows. Pivot Fellowships allow accomplished scientists to explore different areas of mathematics or the natural sciences. The Fellowship enabled Dr. Blitvic to study the connections of seemingly distinct disciplines: She employed probability theory, and statistics to better understand microbes' interactions, growth patterns and movements. You can read more about Dr. Blitvic and her work at the recent newsletter of Simons Foundation.

New interactive wildlife photo-ID game launched with the help of QMUL Centre of Public Engagement

Wed, 27 Aug 2025 23:00:00 +0100

With the help of a QMUL Centre for Public Engagement Large grant, Kostas Papafitsoros recently launched the first of its kind wildlife photo-ID game "TurtleTurtle". The users can play endlessly trying to match sea turtle faces exploiting their unique facial scale patterns. The game draws data from the SeaTurtleID2022 dataset which is the first public large-scale, long-span dataset with sea turtle photographs captured in the wild (Zakynthos Island) originally created to train and test AI-driven automatic photo-ID methods. Several other related research works by Kostas Papafitsoros also inform the game features. There are four difficulty levels for the game as well as a daily challenge. The game is highly suitable for students and wildlife enthusiasts and it has already been warmly welcomed by several sea turtle NGOs worldwide. Give it a try!

New "Statistics and Data Science clinic" launched at the Centre for Probability, ...

Tue, 27 May 2025 23:00:00 +0100

The Centre for Probability, Statistics and Data Science is launching a new "Statistics and Data Science clinic" a free, walk-in statistical consulting service to all members of the broader research community within QMUL. This initiative, funded by QMUL Enhancing Research & Innovation Cultures Fund, aims to provide a range of services, from statistical advice and new methodology development, to discussing potential academic collaborations and possible joint projects across different Schools in QMUL. The idea is to keep the threshold to participation as low as possible: everyone is invited to attend, regardless of their fields and statistical skills. We sincerely hope that this service will help everyone make progress in their research. The Centre will hold two in-person events, 11am-1pm on 9 June and on 23 June in the School of Mathematical Sciences, room MB-502. Refreshments will be available. Although not compulsory, we ask that you register using this form.

Two new members of staff joining the Centre for Probability, Statistics and Data Science

Mon, 26 May 2025 23:00:00 +0100

We welcome Dr Eleni Matechou and Dr Linqi Wang who are joining the Centre as Reader in Statistics and Lecturer in Financial Mathematics respectively.