DURHAM UNIVERSITY

UKRI Gateway to Research · FY 2027 · 2027-12

Submarine landslides can be far larger than any terrestrial landslide and generate exceptionally dangerous and far-travelling tsunami[1-4]. For example, the Storegga Slide that occurred 8,150 years ago offshore Norway involved ~3,000 km3 of material, and is larger than Scotland[5,6] (Fig. 1). The largest terrestrial landslide in the last 350,000 years is ~30 km3[7]. Storegga Slide generated a tsunami reaching heights of up to 20m around many northern European coastlines[3,4]  (Fig. 1). These coasts now host densely-populated cities and critical infrastructure. A repeat of the Storegga landslide-tsunami is probably the most-damaging natural hazard that can potentially impact northern Europe[8], and there is currently no way to warn against landslide-tsunami, unlike earthquake-tsunami. There is a compelling need to understand the frequency and maximum-credible-size of landslide-tsunami from this Norwegian Margin. On average mega-landslides are infrequent, but oceans are warming rapidly at high-latitudes, which may trigger landslides, such as via hydrate dissociation[8,9]. There is a remarkable coincidence between Storegga Slide’s age and the last period of abrupt climate change (‘8.2ka event’) before now[10]. We urgently need to understand if mega-landslides are more common during rapid ocean-warming, and their relationship to glacial-climate cycles[8]. Previously it was assumed a repeat of Storegga Slide needed another ice-stream-advance to deposit more sediment on the shelf-edge, which could fail as another mega-landslide[11]. The final trigger was assumed to be a large earthquake caused by isostatic rebound during deglaciation[11]. This view implies there is just one mega-landslide per ~100,000-year glacial cycle, and another mega-landslide cannot occur in the foreseeable future. However, recent geophysical-mapping of Storegga Slide’s headwall suggested it might be not one but two mega-landslides, with a second mega-landslide occurring ~20,000 years ago[6], well-before deglaciation (‘Nyegga Slide’). Moreover, a newly proposed landslide headwall (‘Solsikke Slide’) suggests another mega-landslide potentially occurred in the last glacial cycle[12], which is ~5 times larger than Storegga[12]. However, ages of Nyegga and Solsikke Slides are poorly known, as neither is penetrated by cores. But if correct, this implies multiple landslides occur in a glacial cycle, and tsunami magnitudes can far exceed that from Storegga. A 20,000-year mega-landslide cannot be triggered during deglaciation, so mega-landslides may not be predictably linked to glacial cycles, casting doubt on conclusions underpinning UK and other national tsunami-hazard assessments. We therefore need to accurately date a longer-record of mega-landslides. Most mega-landslides (e.g. Solsikke Slide) are too deeply buried beneath younger deposits to core and date[12]. But fast-moving mega-landslides that produce tsunami disintegrate to create longer-runout sediment flows called turbidity currents, which produce thick sediment-layers (‘mega-turbidites’) in deep-sea basins[1,8]. Mega-turbidites are an excellent record of mega-landslides as they are less-deeply buried, and easier to date. Storegga Slide produced a 30-m-thick mega-turbidite in the Aegir Basin[5], but there is a thicker mega-turbidite dated to 55-to-60,000-years underneath, and a 10-times-thicker mega-turbidite below that is undated[13]. This project will establish an exact and far longer chronology of mega-landslides over the last ~1.5 million years, to determine their frequency, maximum-credible-size and relationship to glacial-climate cycles. The project is feasible and timely because NERC recently purchased a Giant Piston Corer costing ~£1 million, which can produce 40-m-long cores[14]. This new corer can reach mega-turbidites and some mega-landslides (e.g. Nyegga Slide) that were previously too deeply-buried to core. The project will provide the basis for appropriate responses to mega-landslide tsunami by UK and European coastal flood-risk managers.  

UKRI Gateway to Research · FY 2026 · 2026-09

Magnetic switchbacks are a recently discovered feature of the near-Sun solar wind, the continual stream of plasma and magnetic field emanating from the Sun. Switchbacks are intriguing high velocity folds in the magnetic field where it bends back on itself. Their high number near the Sun suggests that they are a key component of the poorly understood early evolution of the solar wind, but where they come from and how they form is a highly debated active area of research. Following their discovery in 2018 by NASA’s Parker Solar Probe mission, an early suggestion was that they were related to interchange reconnection – the magnetic reconnection of closed and open field lines in the solar corona. The community have since refined this idea, and honed in on coronal jets and other smaller-scale jet-like events as being the most likely interchange reconnection event candidates. Specifically, a newly emerging idea proposed by the Project Lead and others is that MHD waves launched by jets may then evolve into switchbacks as they propagate from the Sun’s atmosphere into the solar wind. The overall goal of this proposal is to test this idea through state-of-the-art MHD simulations validated against satellite observations. This proposal brings together the complimentary skills of Project Lead Wyper, an expert in the MHD modelling of coronal jets, and Co-Lead Matteini and expert in solar wind wave theory and data interpretation. The project builds upon the recent work of Wyper, and will model for the first time the self-consistent evolution of jet-generated MHD waves as they evolve within rapidly expanding and curved solar wind magnetic fields. This will definitively show whether jet-generated waves can evolve into switchbacks, and through forward modelling and validation against satellite data our results will provide constraints for the types of switchbacks formed by this mechanism. This timely project comes at an exciting time for solar and heliophysics with Parker Solar Probe having reached its closest approach where it will now routinely sample the very edge of the solar atmosphere as it transitions into the solar wind. ESA’s Solar Orbiter is about to start moving out of the ecliptic plane to begin sampling solar wind at higher latitudes, and NASA’s PUNCH is about to launch which will remotely image the transition from the Sun’s atmosphere to the solar wind. By linking impulsive coronal reconnection events to their solar wind signatures this project will provide critical modelling needed for understanding and interpreting the results of these missions.

UKRI Gateway to Research · FY 2026 · 2026-09

This project will test the hypothesis that dark matter particles interact with each other, through forces besides gravity. The hypothesis predicts that dark matter's trajectory through collisions will deviate from the purely gravitational trajectories followed by stars. To test it, we will observe astrophysical versions of dark matter colliders, both during a collision and in their long-term steady state. The masses and interaction strengths of Standard Model particles are measured by smashing them together in terrestrial experiments like the Large Hadron Collider. Dark matter cannot be accelerated around Geneva, but it is injected into natural colliders such as the Bullet Cluster. We propose to map the trajectory of dark matter debris from more than an order of magnitude more collisions than previously studied. Our measurements will be interpreted against the latest generation of hydrodynamical simulations, with full hydrodynamical effects and dark matter physics beyond the Standard Model. The proposed test will either confirm a model of 'self-interacting dark matter' (SIDM) on which other measurements are tentatively converging, or rule out dark matter particle interactions as meaningfully influencing the formation of structure in the Universe.

UKRI Gateway to Research · FY 2026 · 2026-09

Ultraluminous X-ray sources (ULXs) are the most voracious accretors we know of, consuming matter at a rate relative to their mass that is far in excess of any other known, persistent systems.  They comprise a compact object, a neutron star or black hole, fed via matter transfer from a companion binary star.  For many years the key question for ULXs was whether they were evidence for a new class of intermediate-mass black holes (with masses between 100 and 100000 times that of the Sun), or whether they instead were stellar-mass black holes (of a few to a few tens of Solar masses) accreting in a rare "super-Eddington" state.  This is where the theoretical Eddington limit, a putative maximum brightness for which the outward pressure of the intense radiation should inhibit further accretion, appears breached.  However, this paradigm changed abruptly a decade ago when we found that several ULXs show pulsations.  These pulsations can only occur in systems harbouring neutron stars, which for their relatively small size (less that 3 Solar masses) must be very extreme accretors, surpassing the nominal Eddington limit by factors up to ~500. The key questions for ULXs now relate to neutron stars - what proportion of the ULX population hosts a neutron star rather than a black hole?  And how is the super-Eddington accretion flow affected by the strong magnetic fields inherent to many neutron stars?  Other important questions yet to be resolved include how does the substantial energy release of ULXs - both from their intense radiation, and carried in the massive energetic winds driven by super-Eddington accretion - affect the galaxies hosting them, particularly at high redshifts where large populations of ULXs are expected?  Here, I present a programme which will exploit my external collaborator status on the eROSITA X-ray telescope to answer these questions. In collaboration with eROSITA consortium colleagues, we have a ULX catalogue based on the first pass of the eROSITA all-sky survey (eRASS1) close to publication.  From this we can confidently predict that the full eRASS:4 data (a combination of all 4 completed sky surveys) will provide a statistically complete and unbiased sample of all eROSITA-detectable ULXs at distances out to 10 Mpc (and beyond that at the more sensitive orbital poles).  We will use a combination of the extensive available archival X-ray data (from XMM-Newton in particular, which already covers > 70% of a pilot sample) and new data to search for signatures of a neutron star in the complete sample of ULXs, including leveraging a new technique we have developed for detecting weak pulsation signals.  This will be used to create the first strong limits on the proportion of the ULX population hosting a neutron star. Any new neutron star ULX detections will be used to deepen our understanding of the accretion flow in such objects, using existing and new data with the best available models to constrain magnetic field strength and accretion geometry.  The study of the compete sample will also extend to creating spectral energy distributions that extend to the infrared and optical wavebands, using the HST and JWST archives to identify counterparts to the ULXs, that can be used to constrain the energy output of a typical ULX, and so to model their effect on galaxy evolution.

UKRI Gateway to Research · FY 2026 · 2026-05

The goal of DNA is twofold. On the one hand, the funds will allow extending the research time of a visit already planned in an existing collaboration between Durham University (DU, United Kingdom), the University of Newcastle (UoN, Australia) and the University of Adelaide (UofA, Australia), funded by the Australian Research Council. On the other hand, the extended visit will be key to form and initiate a new research collaboration in the area of reliability-based design for geotechnical structures involving leading experts in the field from UoN and UofA.

UKRI Gateway to Research · FY 2026 · 2026-03

Progressive Supranuclear Palsy (PSP) is a rare and incurable neurodegenerative disease that affects ~7 in 100,000 people. People with PSP can have many of the same symptoms as people with Parkinson’s disease and their first diagnosis is often Parkinson's disease (PD). On average, it takes about 3 years for a patient to get their correct PSP diagnosis. This delay creates uncertainty for patients and restricts their ability to access appropriate support. Patients also require many additional hospital visits which is costly and burdensome for them and the NHS. Delayed diagnosis is also challenging for developing new treatments, because researchers cannot be certain which patients have PSP.  Patient groups and doctors agree that research into new tools that address the challenge of early and accurate differential diagnosis of PSP and PD should be a high priority. One important difference between PSP and Parkinson's disease is that PSP affects cognitive function early in the disease. PSP patients have often problems with attention, short term memory and planning. Our previous research has shown that people with PSP have particular problems with visual attention, for example when they are trying to find an object that is hidden among distractors (e.g. a ‘Where’s Wally’ type of puzzle). People with Parkinson’s disease do not have this problem. The ability to search for a target among distractors can be measured scientifically using a test called the Bells Test, which test measures how quickly and accurately a person can find and circle bell-shaped objects hidden among other objects. This test is usually used to screen stroke patients for attention problems, but in a recent trial funded by the Vivensa foundation we discovered that people with PSP perform very poorly on this test compared to people with PD. In fact, the Bells test was able to discriminate between people with PSP and PD with specificity of .96. A test with high specificity is particularly useful for rare diseases, as it allows doctors to rule out the most common alternative diagnoses, in this case PD. The aim of this project is to validate the specificity of the Bells test in a new group of 5 people with PSP and 52 people with Parkinson's. Feedback from patients and carers emphasised the importance of accessibility and minimising travel, so our project was designed with the NIHR RRDN as a community based project. RRDN Agile researchers will conduct the tests during a home visit or e-appointment via video-conference. Each participant will be tested twice, once within 4 weeks of enrolling and again after 6 months, which will also allow us to track how their performance changes over time. Faster and more accurate diagnosis of PSP will reduce the costs of multiple hospital visits for carers and the NHS, give certainty to patients and carers, help ensure patients have access to appropriate care pathways and can be a key enabler of trials of new drugs to treat PSP. The Bells test is non-invasive, quick, free and easy to score, so is very economical in terms of financial cost and clinician time. It already widely used in stroke medicine so barriers to adoption are low, and our team includes clinicians who specialise in Parkinson’s disease education for healthcare professionals, so are ideally placed to ensure doctors hear about and use the findings of this research.  

UKRI Gateway to Research · FY 2026 · 2026-03

Context Soil acidification is an urgent and escalating barrier to agricultural productivity in sub-Saharan Africa (SSA). Climate change amplifies nutrient leaching (Ca²?, Mg²?, K?) and the accumulation of toxic ions (H?, Al³?), which degrade soil structure, reduce water retention, and inhibit root growth. Over 35?% of SSA’s cropland (>350 million hectares) is acid-affected (pH < 5.5), especially in East, Central, and Southern Africa. The economic burden is estimated at US?$68?billion annually, equivalent to a ~3% reduction in the region’s agricultural GDP. While conventional solutions such as liming (e.g. calcite or dolomite) are widely used to raise pH to optimal ranges (6.5–7.5), their real-world efficacy is constrained by uncontrolled delivery—heavy rainfall or drought can wash away, over-dissolve, or limit uptake of amendments. In response, embedding dolomite microparticles within a porous, organic polymer matrix offers a route to regulated delivery, simultaneously restoring soil organic matter (SOM). This introduces a rich but unexplored soft matter physics problem. Key themes include: Moisture retention: leveraging polymer physics to link network structure, tortuosity, and water holding Ion transport & reaction: using reaction-diffusion and fluid dynamics to capture ion release, swelling, and internal reactions Soil coupling: applying granular physics and capillarity to model fluid flow, particle rearrangements, and ion transport interactions Challenge Addressed The central challenge is to control the spatial and temporal release of liming ions under realistic soil and weather conditions—minimising loss, overliming, and inefficient uptake. Current approaches rely on trial-and-error, and there is a lack of mechanistic, multi-scale understanding that spans from molecular diffusion within polymers to macro-scale interactions in heterogeneous soils. This gap constrains the development of robust, scalable, low-cost acidification remedies in SSA. Aims & Objectives We propose a physics-driven design and evaluation of composite particles (CPs): dolomite microparticles (DLP) encapsulated in a bio-polymeric matrix (BPM) derived from agrowaste. These materials are affordable, biocompatible, and locally available across Africa. Properly engineered CPs can (i) enhance crop yields by maintaining soil pH, retain moisture and nutrient balance over time; (ii) reduce waste and over liming risks; and (iii) increase soil resilience. The research is structured into 4 Work Packages, with the following objectives: ·      WP1: Create CPs, characterise and model ions transport and water retention as a function of composition and environment. ·      WP2: Quantify effects of CPs on model soil as a function of environmental conditions and density of CPs. ·      WP3: Investigate CP release on a range of real soil samples representative of SSA. ·      WP4: Build an active network enabling the testing, manufacture and deployment of CPs in SSA.   Three PDRAs based in DeKUT (Kenya), Embu (Kenya), and Stellenbosch (South Africa) will carry out experimental, characterization, and modelling tasks, supervised by an international UK-Africa team. The PDRAs will spend 3-6 months in the UK for research, training and cross-team integration. Stakeholder co-design begins with a workshop in Kenya to align technical goals with user needs. Potential Applications & Benefits There are two primary beneficiaries in this project: 1.    The project helps build a pan-African physics-soil network, strengthening capacity in soft matter approaches to soil challenges. 2.    Smallholder farmers across acid-impacted regions in SSA, and potentially other tropical and subtropical regions with acidified soils. Integration with stakeholders such as AGRA, ACTS and local agricultural agencies fosters pathways for scale-up, manufacture, and adoption.  

UKRI Gateway to Research · FY 2026 · 2026-03

The Cambridge spy ring was one of the most audacious and controversial espionage operations of modern history, with implications that upended the mid-century Western intelligence community and that still echo across contemporary British culture and society. With the spy ring instigated in the 1930s by the secret recruitment of Cambridge alumni Guy Burgess, Anthony Blunt, John Cairncross, Kim Philby, and Donald Maclean as long-term intelligence agents by the USSR, over the next decades the career paths of the group went on to deliberately compromise the most sensitive organs of British government and society – including the Security Service MI5, the Secret Intelligence Service MI6, the BBC, the Foreign Office, and even Buckingham Palace – in a way that exploited the unspoken assumptions of class-based loyalty and trust that had previously been central to the operations of the British ‘Establishment’. Crucially, the eventual detection of the spy ring and the defection of Maclean and Burgess in 1951 marked a new phase of their impact rather than an end to their influence, sparking unprecedented waves of debate within British media, scholarship, government, the intelligence community, and the cultural imaginary over the extent of the spy ring and the wider issues of secrecy, loyalty, ideology, and betrayal that their activities laid bare. With 2026 marking seventy-five years since the first public revelation of the spies, this Curiosity project will therefore instigate a multidisciplinary analysis of the complicated legacies of the Cambridge spies across this full and entwined political, social, and cultural landscape. Engaging participants including academics, archivists, cultural producers, educators, and government stakeholders, the project’s aim is to address a series of interlocked research areas, which include: ·         interrogating the legacy of these spies within the practices of government and the intelligence community, and assessing the lessons they might still have for current security and intelligence contexts;     ·         understanding how the ring influenced cultural and media understandings of espionage, intelligence, communism, and the conduct of the Cold War; and ·         analysing how the evolving revelations about the Cambridge Five impacted wider societal debates over trust, loyalty, and political ideology, in these cases further complicated by questions over the spies’ sexuality, class, and the operation of ‘the Establishment’. To address these questions, the objective of this project is to deliver a programme of activities (including interdisciplinary research events, stakeholder workshops, engagement with writers and educators, and public-facing talks) that will generate research publications, digital resources, and pedagogical material. Overall, through this, our project will lead to new understandings of the spies themselves and how these themes have shaped modern post-war Britain, and more broadly will establish new frameworks for understanding other evolving debates at the interface of the secret state and British society.

UKRI Gateway to Research · FY 2026 · 2026-01

The PolyForm team are working to transform the environmental footprint of an essential human activity - the routine cleaning of clothes and other textiles. In the UK alone, dishwashers and washing machines use 360 billion litres of water each year and cost households £1.6bn in energy bills, with the majority of the electricity consumed in heating water (Energy Saving Trust, 2013). High-performance laundry and fabric care products enable effective cleaning at low temperatures with minimal quantities of water, and extend garment lifetimes, avoiding their disposal in landfill sites. The performance of these formulations is only possible through the use of synthetic polymer additives, primarily derived from petroleum-based sources. Unfortunately, these often persist in the environment over long timescales, greatly reducing the potential environmental benefits presented by their use. Replacing current technologies with biodegradable alternatives gives us an opportunity to drastically improve the environmental profile of many essential consumer products, but achieving this goal is complicated by the scale of the challenge: requiring not only the development of new sustainable materials, but ensuring their functionality in complex, multi- component formulations. Whilst there has been significant drive to develop biodegradable polymers, and build an understanding of the mechanisms by which they are removed from the environment, there remains a substantial knowledge-gap which renders the prediction of a material's biodegradability challenging. This difficulty presents a major bottleneck in the process of reformulation. PolyForm presents a unified approach to the replacement of synthetic polymer additives in laundry and fabric care formulations, working in close partnership with industry leaders Procter and Gamble to fully address the scope of the challenge. PolyForm is delivering biosourced polymer additives capable of high levels of functionality within detergent formulations, and elucidating the structure-function relationships contributing to performance though detailed mechanistic investigations, in order to enable improved design. The PolyForm team are simultaneously developing new models and tools to assess and predict the biodegradability of new and existing materials, streamlining the process of incorporation of high- performance biodegradable additives within consumer products. These breakthroughs will have impacts far beyond the arena of laundry and fabric care.

UKRI Gateway to Research · FY 2026 · 2026-01

Most living organisms carry DNA called transposable elements (TEs or transposons), which were once foreign to organisms but are now part of their DNA. These TEs originally came from ancient viruses and are best known for their ability to move their genomic locations and thereby cause gene mutations. In some crop plants, they can make up about 90% of the total DNA, and thus pose a greater threat to plant DNA stability and consequently crop production. Because of their mutagenic nature, they are tightly controlled by the host through the so-called epigenetic silencing mechanisms. These involve DNA methylation and heterochromatin formation, which are primarily guided by small interfering RNAs (siRNAs). In plants, there are two types of siRNAs: longer siRNAs of 24 nucleotides that are generated at already methylated DNAs and act to further reinforce the silent state (a process of which is called RNA-directed DNA methylation, in short RdDM), and shorter siRNAs of 21 and 22 nucleotides that establish new DNA methylation at naïve and active TEs. Intriguingly, these shorter siRNAs are produced exclusively by TEs and not by regular genes, however, the mechanisms underlying its selective action towards non-native DNA remain largely unknown. Therefore, this research project aims at unveiling the plant’s strategy to detect foreign DNA and initiate the epigenetic silencing, preventing them from damaging the host genomes. We previously found that TE RNAs are weak in translation, which can then lead to RNA localisation to cytoplasmic compartments known as stress granules, where the siRNA biogenesis pathway exists. Of note, reduced translation is frequently associated with RNA cleavage, which is a critical requirement for the siRNA biogenesis factors to act upon. However, the cellular mechanism linking translational inefficiency and entrance to the siRNA biogenesis pathway has been entirely unknown. To fill this gap, we paid attention to the No-go RNA decay (NGD) pathway, an under-investigated RNA quality control system in plants, which senses the stalled ribosome and triggers RNA cleavage. Importantly, our pilot experiments revealed that the key NGD components are localised in stress granules, and the loss of NGD results in compromised siRNA production. Through this project, we aim to further dissect the role of NGD in the initiation of epigenetic silencing in Arabidopsis, by 1) investigating the physical association of NGD with RdDM, 2) profiling the siRNA and RNA cleavage landscapes in the NGD mutants, and 3) characterising the direct RNA targets of NGD. Understanding the initiation of epigenetic silencing can have broader impact beyond transposon control and genome surveillance; for example, the shorter type of siRNAs (21 and 22 nucleotides) can be generated when plant cells are infected by viruses and express foreign transgenes. This implies that the plant siRNA-mediated epigenetic silencing pathway is an innate defence mechanism of the host against non-native DNA. Therefore, this research project will help improve antiviral treatments for plants and innovate the plant synthetic biology platforms. Overall, this work is relevant to the BBSRC research priority in “understanding the rules of life” as well as this year's BBSRC spotlight in “plant health”.

UKRI Gateway to Research · FY 2025 · 2025-12

Polymers are assemblies of large numbers of molecules linked together by chemical bonds. From a theoretical perspective it is natural to ask if the properties of systems of polymers can be predicted from first principles. That is, if you are told the properties of the molecules, can you deduce the properties of the polymers the molecules can form? Solving this problem with full physical fidelity is extremely difficult. Fortunately, for many important questions, simplified mathematical models produce meaningful and practically useful answers. One type of question that can be asked about polymers concerns the existence of a gelation transition. This transition is familiar from everyday life: when preparing a gelatine dessert, hot water is combined with a concentrated form of gelatine. The result is a liquid that, once cooled, sets into a firm gel. Physically, at high temperatures there are few chemical bonds, and as a result, the molecules form only small polymers that disperse throughout the water. At low temperatures chemical bonds form more readily, and a dense network of them produce a global structure: the gel. This proposal is about a model for the gelation transition of branched polymers, which occur when the molecular structure prevents chemical bonds from forming cycles. The constraint defining branched polymers leads to many fascinating connections. Namely, questions about branched polymers can be reformulated into questions about a mathematically well-formulated quantum field theory. Quantum field theory is a notoriously challenging subject from the mathematical point of view. Recently, however, it has been realised that it is possible to utilise the field theory perspective to show that branched polymers cannot form a very thin layer of gel: there is no gelation transition in two dimensions. There is, however, a gelation transition in three dimensions! Many questions about branched polymers remain unanswered, closely mirroring our mathematically incomplete understanding of quantum field theory. The first goal of this proposal is to develop our understanding of the absence of a gelation transition in two dimensions. This is known as asymptotic freedom, an important feature of our most fundamental physical theories. A thorough mathematical understanding is lacking, and this proposal will make progress by developing our understanding in the branched polymer context. The second goal of this proposal is to understand the fact that branched polymers exhibit self-organised criticality when a gelation transition does occur. While this is understood in specific contexts from a probabilistic point of view, it is not understood from the field theory perspective. The study of branched polymers is rich as it lies at the crossroads of many mathematical subjects, and hence branched polymers provide a concrete setting in which to develop tools of much wider applicability. In particular, this proposal will develop methods for studying coagulation-fragmentation dynamics as well as rigorous renormalisation group methods. These are topics of broad theoretical and applied interest, and it is expected that their development will lead to future applications in probability theory and theoretical computer science.

UKRI Gateway to Research · FY 2025 · 2025-12

Organofluorine chemistry is critical to our modern lives with many of the everyday drugs and materials we take for granted containing carbon-fluorine bonds. Therefore, investment into the development of new fluorination technologies is vital to secure ready access to societally and economically critical compounds whilst offering avenues to protect UK jobs that could be jeopardised by upcoming environmental restrictions. The Selective Elemental Fluorination Future Leaders Fellowship (SELF-FLF) is an academic/industrial collaborative programme to leverage the most atom-efficient, cheapest, and industrially applicable source of electrophilic fluorine atoms, elemental fluorine (F2), to develop a suite of new selective fluorination processes. It will tackle the pressing need for new fluorination approaches whilst minimising the generation of toxic and biopersistent perfluoroalkyl substances (PFAS). It will facilitate a dramatic step change in how fluorination is conducted at scale and will allow UK industry to create new and improved products to positively impact the UK’s economy and society. Fluorination is vital to many industries, for example, >20% of small-molecule drugs and >25% of agrochemicals incorporate fluorine atoms whilst the pharmaceuticals (£40.7B), plastic and rubber products (£27.7B) and electronics (£25.3B) sectors all rely heavily on fluorinated compounds. However, there are two main hurdles currently facing industrial fluorination. Firstly, selective fluorination of molecules relies on expensive and highly atom-inefficient reagents. This means industrial-scale selective fluorination can be extremely challenging and prohibitively expensive. Secondly, there are environmental implications for PFAS generation, which are a class of molecules that are highly biopersistent and harmful to the environment. Due to this, PFAS are about to see a high degree of restriction in their manufacture and application through governmental legislation (e.g. EU REACH programme). PFAS therefore need to be replaced, minimised as byproducts in fluorination reactions, and the manufacturing sites for them repurposed for other less polluting selective fluorination processes. The objectives of SELF-FLF are to: Develop new fluorination processes using F2 in combination with substrate activation methods, which are atom-efficient, economical, scalable and industrially relevant to safely and cleanly generate vital fluorinated products. Work closely with the SELF-FLF industrial partner F2 Chemicals (the largest manufacturer and user of F2 in the UK) to commercialise the developed fluorination processes. This will mitigate against upcoming environmental restrictions, thus increasing job protection in UK chemical industry and economic prosperity. Develop the SELF-FLF team and corresponding F2 facilities to provide unique opportunities for innovation within the UK industrial landscape. This will establish Durham as a centre for fluorination technologies and one of the few academic sites globally capable of developing elemental fluorination reactions. Secure the training pipeline of fluorine chemists for the UK chemical industry as well as develop a unique skillset within the PI. This will facilitate the PI in becoming a world leader in elemental fluorination. The SELF-FLF team will bring together a range of expertise as well as unique equipment only available at Durham University to deliver an innovative, ground-breaking programme that will deliver substantial benefits to the PI, Durham University, UK PLC, and the UK taxpayer offering clear pathways to commercialisation and wider impact for both chemical industry and society.  

UKRI Gateway to Research · FY 2025 · 2025-12

The UK’s targets for Net Zero are ambitious but must be realised if we are to prevent the worst effects of climate change over the coming decades. A key component of Net Zero is the conversion of sources of electricity to those obtained without the use of fossil fuels and for the UK this means to a large extent offshore wind. The UK has massive potential wind energy resources some of which has been realised already to an impressive extent. We are all now familiar with large offshore wind farms around our coasts and to an extent the UK has been a leader in this area with other countries starting to ramp up their own offshore wind developments (e.g. China and the USA). Offshore energy generation (including wind but int he future potentially tidal, current and wave energy could become significant) requires considerable infrastructure to be developed, maintaining and decommissioned for a resilient energy delivery service. At present, most offshore wind turbines are supported on monopile foundations driven into the seabed, but already we are seeing the advent of floating offshore wind where superstructures are tethered to the seabed via cables and anchors of various different types. As well as the foundations, the electrical services needed to connect a generation source to the grid involves major construction works, e.g. the laying in trenches of cables. As activity increases and time goes on the industry is faced with perhaps new issues, e.g. of having to build new infrastructure adjacent to existing, and the whole issue of decommissioning installations that are obsolete or need replacement. We cannot create a situation where old windfarm locations become areas of brownfield seabed. This sets the scene for this particular proposal which seeks to improve the computational tools that engineers can use to design new installations of the type described above, simulate various scenarios to assess behaviour over time and make predictions of the effect of operations (installation or decommissioning) on the seabed flora and fauna. Practical tools have to be computational. Field testing is expensive, dangerous and site-specific; physical modelling in the laboratory is useful but many issues exist with fidelity to a real site. Computational modelling, properly validated against field data and the lab testing mentioned provides the only way to make the predictions the industry will need to contribute to the Net Zero goals as outlined above. The proposed project starts from a computational software developed at Durham University, which to date has been a vibrant source of solutions to interesting research questions through three EPSRC-funded projects, and will take it to the next level as regards utility to the industry tasked with new offshore developments. The project has a suite of outstanding industry supporters who will guide the researchers during and after the project to ensure that the outputs are indeed the tools that industry needs.

UKRI Gateway to Research · FY 2025 · 2025-11

Context Cytokinesis, the division of one cell into two, is crucial for an animal’s development and healthy life. Errors in this process can lead to cell death or mutations. Research on cytokinesis has largely focused on simple, single cell models systems, and has led to the identification of key proteins and processes. However, emerging evidence suggests that some cell types show variations from these models and clinical studies have further demonstrated this variability, and the limitations of our current models. Challenge In this wider context of reassessing our understanding of cytokinesis, we must now use biological model systems that allow us to demonstrate how cytokinesis may vary between different types of cells, and where the effects of cell environments can also be recognised and measured. To address this challenge, we have established embryos from the round worm, Caenorhabditis elegans, as an ideal tool to study cytokinesis in a complex environment. We have previously shown that specific cells in the embryo can divide successfully despite depletion of what were previously thought to be essential proteins (f-actin polymerase and f-actin).  This successful division requires signalling from an adjacent cell and our preliminary data indicate the involvement of the ‘Wnt pathway’ a cell signalling pathway important for cell fate determination during development and frequently goes wrong in tumours. In this project, we seek to understand Wnt’s ability to promote successful cytokinesis in neighbouring cells. Aims and Objectives We aim to understand the molecular requirements for cytokinesis using the C. elegans model system, by manipulation of signalling pathways and tracking of cellular components in real-time during division. Our preliminary data indicate that Wnt signalling controls the position of the ‘mitotic spindle,’ a key cellular structure that signals where the cell will divide, and that this impacts other cytokinesis components. We will investigate the role of Wnt signalling in promoting cytokinesis with three specific objectives: To dissect the ability of Wnt signalling to enable successful EMS cytokinesis. To determine how Wnt signalling regulates spindle position during cell division. To identify how Wnt regulation of spindle position enables cytokinesis. Potential Applications and Benefits Cytokinesis is a central biological process, and its regulation by contextual factors such as cellular signalling is crucial to understanding the complexity of cell division in multicellular organisms. By elucidating how Wnt signalling promotes successful cytokinesis, this project will advance the paradigm of context-specific cell biology, showing that even fundamental processes like cytokinesis can vary depending on the cellular environment. These insights have significant implications for understanding diseases, such as cancer, where both cytokinesis and signalling pathways are often dysregulated. Relevance to BBSRC’s Long-Term Research and Innovation Priorities This project aligns with the BBSRC’s objective of advancing the frontiers of bioscience discovery by expanding our understanding of the rules of life. By integrating knowledge of cell signalling with the mechanics of cell division, it will deliver an integrated understanding of health. Insights gained from this work may also inform therapeutic strategies for diseases characterized by dysregulation of essential biological processes.

UKRI Gateway to Research · FY 2025 · 2025-11

Many rivers, including >50% of UK channels, have gravel beds. Managing these channels requires predictions of when, and how much of, the riverbed will move, for applications including designing infrastructure, mitigating flood risk and river restoration. The amount of sediment transport depends on the balance between the shear stress applied to the bed by the flow (t), and the critical shear stress (tc) at which riverbed sediment starts to move. tc is strongly affected by grain size, but large uncertainties in predicting tc remain. These uncertainties in tc are problematic because they mean that we currently cannot predict bedload transport to better than an order-of-magnitude. One important factor that controls tc is how grains are arranged on the riverbed. The project team has led recent advances in understanding these controls, developing new methods to measure and model the impact of grain arrangement, which has produced a paradigm shift in predicting tc. However, a substantial omission so far, and from all bedload transport models, is that they neglect the effects of commonly occurring cohesive material that sticks gravel together. This includes clay (often alongside sand and silt), biological secretions such as Extracellular Polymeric Substances (EPS) produced by bacteria and biofilms, and silk threads produced by larger organisms such as mussels and caddisfly larvae. Previous work has shown that such cohesive material can increase tc by up to ten times, and so understanding its effect is essential for improving bedload predictions. However, the current data are derived from varying methods, and consider different cohesive materials individually rather than in combination. Consequently, they do not provide a comprehensive understanding of cohesive effects, nor are they sufficient for model development. Our aim is to incorporate cohesive effects into our models for tc, improving their predictive capability and making them widely applicable. To achieve this, we will combine field data, laboratory flume modelling, and theoretical model development, and will address both non-biological (clay) and biological (EPS and caddisfly silk) cohesion. Our objectives are: O1) To monitor where and when cohesive material is found in gravel bed rivers, creating the first dataset that includes both biological and non-biological cohesion. This will identify where cohesive material is most prevalent in river channels, and inform conditions for flume experiments in O2. O2) To measure how biological and non-biological cohesive materials, alone and in combination, affect resisting forces, sediment structure, velocity profiles and tc in gravel-bed channels. By measuring all of these effects concurrently our dataset will be the first that is sufficient for model development. O3) To use novel flume experiments with see-through beds to measure how cohesive material affects near- and within-bed turbulence and velocity profiles. These experiments supplement O2 by quantifying near-/within-bed flow that cannot be measured in O2, and will be the first application of this technique to beds with realistic gravel-shaped grains. O4) To use the outputs from O1 to O3 to develop and test a new version of our entrainment model that will incorporate cohesive effects. Our findings will improve predictions of tc and bedload in gravel-bed rivers with cohesive material. Our outputs will range from rule-of-thumb recommendations for situations without field data, to methods for parameterising our entrainment model when it is possible to measure cohesive material. These will be of benefit to anyone involved in managing and studying gravel-bed rivers.

UKRI Gateway to Research · FY 2025 · 2025-10

Ultracold molecules have emerged as an exciting and powerful platform for experiments in quantum science. The molecules have a rich internal structure of long-lived rotational, vibrational and spin states, and there are controllable long-range dipole-dipole interactions between them. This has motivated many proposed applications spanning ultracold chemistry, precision measurement, quantum computing and quantum simulation of many-body phenomena. The field is flourishing: recent highlights include the control of ultracold molecular collisions, the creation of quantum-degenerate Bose and Fermi gases, and the first entangling gates between molecules. Looking to the future, there is a need to bring new types of molecules into the ultracold regime. Currently, two types of ultracold polar molecules are prevalent: bialkali molecules, such as KRb and RbCs, that can be assembled from ultracold atoms, and free radicals, such as CaF and SrF, with properties that make direct laser cooling feasible. Bialkali molecules are produced at high densities, leveraging the powerful cooling of atoms. They do not possess an electronic magnetic moment. In contrast, laser-cooled molecules possess both an electronic magnetic moment and an electric dipole moment, but are produced at low densities. For many applications it is desirable to prepare the molecules in an optical lattice with a high fraction of the sites containing a single molecule. This has been achieved for bialkali molecules by preparing atom pairs in the lattice prior to forming the molecules. However, for simulating lattice spin models relevant to quantum magnetism, it is desirable to use molecules with both magnetic and electric dipoles. The low densities of laser-cooled molecules currently preclude this. We will address this challenge by extending the association technique that has proved so successful for bialkali molecules to a new class of molecules. Specifically, we will associate alkali-metal atoms (Cs) and alkaline-earth-like atoms (Yb) prepared in an optical lattice to form CsYb molecules possessing both magnetic and electric dipoles. Previous attempts to produce molecules of this type from ground-state atoms have been thwarted because the scattering resonances needed for the association are extremely narrow. However, our preparatory work has identified resonances involving the metastable clock states of Yb that are suitable for molecule formation. To find and successfully exploit these resonances for molecule formation requires a research programme integrating theory and experiment at all stages. Our approach blends established techniques from the quantum-gas, optical-lattice-clock and ultracold-molecule communities to: (1) Prepare Cs-Yb atom pairs in optical lattices. (2) Access scattering resonances using the metastable states in Yb and form molecules using magnetoassociation. (3) Stabilise the molecules by transferring them to the ground electronic state and then using stimulated Raman adiabatic passage to the lowest rovibrational energy level. The combination of Cs and Yb is particularly advantageous. Yb offers 6 different isotopes (bosons and fermions) with usable natural abundances. When combined with Cs this allows significant tuning of the reduced mass and hence the interspecies scattering properties. A priori, we do not know which isotope will be the most favourable and will therefore explore several, developing and refining the underpinning theory as we progress. Our success will usher in a new era for quantum simulation using ultracold molecules and our application of optical-clock technology to a molecular system will opening exciting new avenues for probing fundamental physics.

UKRI Gateway to Research · FY 2025 · 2025-10

In seed plants, wood provides both the mechanical strength that enables plants to grow tall and a means to transport water to the canopy. Wood is derived from cell divisions in a meristem referred to as the cambium. Although most obvious in forest trees, the cambium is largely conserved in seed plants, including in small rapid-cycling annuals such as Arabidopsis. Arabidopsis is a powerful model organism in which to study cambial regulation due to its rapid-cycling and the vast genetic resources available. Remarkably, the stem cell factors that underpin cambium function were unknown, but we have recently shown that members of the PLT and ANT family of transcription factors represent these elusive stem cell factors1. Our experiments demonstrated the expression of PLT and ANT family members are regulated by a plasma membrane localised receptor kinase, PXY, and its cognate ligand TDIF. However, TDIF-PXY has a suite of additional target transcription factors that perform other roles in cambium maintenance including regulating cell division (WOX4, WOX14)2,3, contributing to patterning (TMO6, LBD4)4 and repressing xylem differentiation (BES1)5. It is not known how the newly identified PLT and ANT stem cell factors integrate into this broader system, but understanding this is a prerequisite to understanding cambium function. The knowledge gap is addressed within this proposal. Given the complexity of the system, to understand how the newly identified factors integrate into the previously known cambium regulatory network requires transformative technologies, in the form of mathematical modelling, to be coupled with genetic approaches. Previous mathematical models of cambial development have captured elements of pattern but not growth1,6 or have incorporated growth but been unable to capture pattern without including hypothetical factors and discarding an important driver of cambium cell division, WOX47. The model proposed here will bridge this modelling gap. We have discovered a new set of transcription factors, members of the PLINC family, that our preliminary unpublished data suggests repress the expression of some TDIF-PXY regulated genes. We hypothesise that a mathematical model that incorporates growth without hypothetical factors and includes WOX4 and members of the PLT and ANT families will pattern correctly upon incorporation of these PLINC transcription factors. Biological data will be used to parameterise the model and model outputs will be tested in planta. Our proposal is focused on generating new fundamental knowledge aimed at understanding the rules of life. Nevertheless, we have previously demonstrated that TDIF-PXY signalling is conserved in forest trees so the discoveries made here will form the basis of future applications. Understanding cambium function is critical because wood represents a globally significant carbon sink and a source of renewable biomaterials. As such the research programme described here could contribute to tackling strategic challenges around renewable resources and clean growth in future.

UKRI Gateway to Research · FY 2025 · 2025-10

The central requirement of project safety, stability, and resilience of complex underground systems leads to demands for more efficient computational modelling tools to assist design and decision-making during the project life cycle. The concept of Digital Twins (DTs) provides a robust solution to monitor a construction project during its life cycle, predict its behaviour based on integrated holistic computational models, and protect it from hazards by virtually controlling the physical processes with its digital counterpart. Leveraging the power of a computational framework based on CutFEM combined with a BIM platform incorporating CAD-based data, the TwinSSI project will develop a comprehensive DT for underground design and construction. To validate the computational framework, real-scale experiments of tunnel-soil-structure interaction will be performed at the National Buried Infrastructure Facility (NBIF) at UoB. Moreover, the developed DTs will be applied to real case studies co-created with the industrial partners Network Rail and Maidl Tunnel consultants. The TwinSSI project will thus, for the first time, create and validate detailed DTs in the domain of soil-structure interaction modelling. The project outcomes will lead to a new paradigm for project planning and monitoring by geotechnical engineers.

UKRI Gateway to Research · FY 2025 · 2025-10

Nowadays, comfort is a design requirement of all structural products that guarantees the quality and competitiveness. Noise constitutes a significant form of environmental pollution that impacts the lives of hundreds of millions of people globally, leading to various socio-economic consequences. Intense vibration has the potential to jeopardise both the structural integrity and the performance of equipment and hardware, and produce significant level of noise thereby affecting the comfort of individuals in several aspects. Subways represent a primary source of ground-borne noise and vibration in urban areas. Over 7.5 million Europeans face potential disturbance from railway noise and vibrations. In response to public concerns, governments have established laws and regulations to limit the permissible exposure of citizens and facilities to ground-borne noise and vibration. The goal of the META-NOVIB project is to develop a comprehensive framework to effectively predict and control the vibration and noise induced by underground railway tunnels using digital twin technology supported by machine learning tools. This system provides valuable insights for engineering decisions throughout the operation and maintenance of these tunnels. Additionally, it evaluates the performance of seismic metamaterials (SMM) in attenuating the level of noise and vibration to meet the allowable limit. META-NOVIB will provide an integrated platform for visualisation and real-time prediction and virtual control of the railway-induced noise and vibration during the operation and the maintenance phase. Thus, the output will have wide implications on the health of nearby residents due to vibrations and prevent any structural damage to historical buildings or structures, with high academic and industrial impact.

UKRI Gateway to Research · FY 2025 · 2025-09

Afghanistan's post-2001 elections were meant to swap bullets for ballots. Yet each presidential vote provoked new disputes, shifting politico-military alliances and, in 2021, the collapse of the political order built after the international intervention. This fellowship asks why those polls failed to deliver the stability and legitimacy donors expected and shows, through this cautionary case, what external interventions should not do. My doctoral research is grounded in more than fifty confidential interviews with Afghan presidential candidates, opposition figures and other key power-brokers. Their accounts point to a clear pattern: in war-torn and divided societies, elections are not one-off gateways to democracy but recurring bargaining rounds in which powerful actors renegotiate access to state resources and government positions. The relative bargaining power of each electoral camp—set during pre-election coalition formation and endorsement deals—largely shapes what happens on election day and afterwards. When an incumbent and external sponsors make their backing crystal clear, rival elites accept side-payments and step back; when that backing is ambiguous, violence—or the threat of it—is used to secure a better deal. Grasping this sequence is essential if future peace-support missions are to avoid Afghanistan's fate. During the twelve-month fellowship, I will translate these findings into outputs that serve scholars, policymakers and the Afghan public. First, I will write a book, The Power of Bargaining: Elections, Rents and War-to-Peace Transitions in Afghanistan, and publish a peer-reviewed article comparing elections in the midst of war. Second, I will prepare a concise policy brief and host a practitioner workshop for officials in the Foreign, Commonwealth and Development Office, relevant UN agencies and democracy-assistance NGOs, distilling how the configuration of pre-election elite coalitions—and the external signals that inflate or shrink their holding-power gap—rather than election-day logistics or anti-fraud technology, determines whether a vote ends in swift concession or in months of destabilising dispute. Third, I will script and release five short video explainers in Dari (Afghan Persian) and Pashto so that Afghans inside and outside the country can access—and debate—the research in their own languages. The fellowship is hosted by Durham University's Global Security Institute under the mentorship of Professor Roger Mac Ginty, a leading scholar-practitioner in peacebuilding whose networks link research to decision-makers. Durham's advanced training programme will deepen my skills in causal inference and in translating research for policy professionals, whilst a light teaching load (no more than six hours per week) will strengthen my publication record and classroom experience. By reframing elections as iterative power bargains rather than democratic finish lines, the project sharpens debates on democratic resilience, state fragility and external intervention. It offers donors concrete guidance on how pre-election coalition deals and external endorsements condition the risk of violence, whilst giving Afghans—journalists, students and activists—the tools to scrutinise future political openings more effectively.

UKRI Gateway to Research · FY 2025 · 2025-09

My research examines the following questions. First, why and how do people become attached to fossil-fuelled ways of life in a wider context where greenhouse gas emissions urgently need to be cut? Second, can attachments to fossil-fuelled lives be disrupted? If so, how? I explored these questions in my PhD. I examined a case study of a northern English town – one which found itself at the centre of controversy because of plans to open a coal mine in the area. I showed that many of the town’s residents re-attached to the promise of new coal extraction, even though people had detached from coal industries following the closure of coal mines in the 1980s. I argued that this signalled a warning about how fossil fuel attachments can be unexpectedly resurrected, even when a transition towards low-carbon energy production has happened. At the same time, I showed how a small group of residents in the area attached to the possibility of a ‘net zero’ transition, where society is no longer dependent on fossil fuels. They did so after taking part in a democratic citizens’ forum, learning about climate change and devising solutions to reduce greenhouse gas emissions. I argued that this signalled an opportunity: That attachments to forms of life which are no longer reliant on fossil fuels can be formed, often at speed, when people re-imagine alternative futures. In this fellowship, I have two key objectives – to communicate the findings of my doctoral research to new audiences, and to open opportunities to carry out new research about people’s attachments to fossil fuelled and post-fossil fuelled ways of life. First, I aim to disseminate my PhD findings among two audiences – academic researchers who examine the politics of energy transitions, and climate change policymakers and campaigners. To publicise my research to academic scholars, I am: publishing my doctoral thesis as a book; writing a new paper in an academic journal, based on one of my thesis chapters; and giving presentations at two academic conferences. Policymakers and campaigners typically engage with content which is shorter than academic publications. I am writing two articles for online media and speaking about my research on a climate change podcast. In addition, I am co-producing a policy brief based on my, and other comparable, research. The brief will be shared with policy organisations (from local authorities and national government audiences to think-tanks and political parties). Second, I am applying for new funding opportunities to build and expand on my research. I will propose examining cases where workers detach from high-carbon industries (in the oil and gas sector) to attach to sustainable industries (e.g., work in renewable industries). My focus is the politics of their actions – how they challenge the power of fossil fuel industries by pointing to the possibility of alternative futures, even when doing so involves embracing risks and uncertainty. Overall, I want to use my research to contribute to tackling the climate emergency, so that we can protect people and future generations. My hope is that my research can play a role in contributing towards that outcome.

UKRI Gateway to Research · FY 2025 · 2025-09

The N8 centre of Computationally Intensive Research (N8 CIR) supports research communities across the North of England with the aim of advancing the art of the possible through advanced computational and data-intensive practice. Through the provision of the EPSRC Tier 2 High Performance Computing (HPC) system, Bede, N8 CIR also serves a national mission with a focus on enabling capability accelerated workloads for both Artificial Intelligence (AI) and simulations. Bede’s unique architecture supports coherence between the CPU and GPU, allowing larger problems to be tackled than those which fit within a single accelerator card, through Grace-Hopper and Power9/V100 technologies. Working in conjunction with the N8 CIR, interventions are proposed that aim to catalyse trust and confidence in AI capabilities within the Bede user community by providing a safe-space and expertise to trial AI approaches, enhancing scientific method alongside critical thinking, respectful of the sensitivity of information submitted to AI tools. These include: The identification of and implementation of the hosting of key models and data applicable to the current Bede community to enhance AI for science. Digital Research Technology Professional (dRTP) resource to collaborate with researchers in the application of the deployed models/data to act as exemplars to domains. dRTP resource to design and expose custom interfaces to these models, in conjunction with researchers to improve the accessibility of AI tools on HPC to non-HPC users, such as parts of workflows for bench scientists. To document and share practice including the production of case studies that will be published via the N8 CIR website and presented at National events such as Computing Insight UK (CIUK) and Durham HPC days. The adoption of AI methods has widespread transformative potential for scientific discovery. Through existing mechanisms of the N8 CIR, application of interventions with transferability to other platforms, breadth of beneficiaries (both within and beyond the Bede user community) and of high strategic value to stakeholders will be targeted.  

UKRI Gateway to Research · FY 2025 · 2025-09

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

UKRI Gateway to Research · FY 2025 · 2025-09

Doctoral Training Partnerships: a range of postgraduate training is funded by the Research Councils. For information on current funding routes, see the common terminology at https://www.ukri.org/apply-for-funding/how-we-fund-studentships/. Training grants may be to one organisation or to a consortia of research organisations. This portal will show the lead organisation only.

UKRI Gateway to Research · FY 2025 · 2025-09

Advances in sensing technology have made it possible to collect large volumes of time-series data on many variables. In a diverse array of fields, a key question is whether such data can be used to learn directed relationships between variables. In other words, whether changes in one variable consistently precede those in another, an idea formalized in the concept of ‘Granger causality’. In this project, we use two motivating examples from neuroscience, where learning the drivers of change could lead to substantial improvements in our understanding of the underpinning physiology. The first concerns abnormal brain activity patterns which typically affect people with epilepsy. The second concerns the 24-hour biological clock, measured by a range of physiological variables. Graphical vector autoregressions (VARs) are a popular tool for learning such lag-lead relationships in multivariate time-series. A VAR of order p expresses the observation at time t as a regression on the preceding p terms. The pattern of zeros in the autoregressive coefficients has a graphical interpretation: absence of an edge from variable i to variable j is tantamount to i being Granger non-causal for j. In the Bayesian inferential framework, this kind of sparsity in model parameters can be accommodated by prior distributions which assign non-zero probability to every pattern of zeros. One limitation with current approaches to fitting graphical VARs is that useful Markov properties of Granger causality graphs rely on the process being stable (e.g. constant mean, variance, covariances), at least locally. For the process to be stable, the autoregressive coefficients must lie in a constrained space with a complex geometry, and so stability is generally assumed without being enforced. This can be problematic when there are not enough data to learn, with certainty, that a process is stable. A second limitation arises when data, though recorded at discrete intervals in time, would be more naturally described through the underlying continuous-time process. In this case, although the time discretisation is often chosen for convenience, the notion of Granger causality is dependent on it. Directly modelling the continuous-time system through the continuous-time analogue of a sparse VAR overcomes this problem but, again, enforcing stability imposes complex constraints on the parameters. Our aims are therefore: Develop prior distributions for VARs and continuous-time linear systems that simultaneously encourage sparsity in the parameters and enforce stability; Develop associated procedures for computational inference; Apply our ideas to the motivating applications from neuroscience.  There are two main challenges. The first is that the complex constraints imposed by the stability condition make it difficult to specify a prior distribution with appropriately restricted support. This problem is then compounded by requiring the prior to be sparse, essentially making its dimension one of the unknowns. There is then a second challenge: constructing a Markov chain Monte Carlo (MCMC) sampler over a constrained space of unknown dimension. In our two applications, interpretation of Granger causality networks has huge potential to improve understanding of the relationships between variables, with important implications for the treatment of disease. There are also numerous other fields which rely on network inference in their research and which therefore stand to gain from our methods. This includes genetics, finance and microbial ecology, with potential benefits further down the line from the broader agenda their research serves, be that society, the economy or the environment.