Schedule

0900	Registration and breakfast	Great Hall
0930	Welcome	Fellows' auditorium
1000	Dr Lily Whitler
1045	Break and Posters (Group A)	Great Hall
1115	Jake Stuchbury-Wass	Fellows' auditorium
1135	Aparna Kaaraal Mohan
1155	Juliana Ferraro
1215	Dr Peter Pihlmann Pedersen
1300	Group photo
1300	Lunch and Posters	Great Hall
1400	Jieni Wang	Fellows' auditorium
1420	Eleni Papafilippou
1445	Prof. Louise Hirst
1530	Break and Posters (Group B)	Great Hall
1600	Prizes, closing	Fellows' auditorium
1845	Drinks reception (by invitation)	Fellows' Dining Room
1930	Dinner (by invitation)

Invited talks

Dr Lily Whitler

Dr Whitler is the Gavin Boyle Fellow at the Kavli Institute for Cosmology at the University of Cambridge. She is an observational astrophysicist, and her research focuses on understanding the earliest generations of galaxies and the process of cosmic reionisation.

New Eyes on the Sky: Insights into the Early Universe from James Webb Space Telescope Observations

The James Webb Space Telescope (JWST), NASA's current flagship observatory, was launched in December of 2021 and has revolutionized the field of extragalactic astronomy in only a few short years. JWST's unprecedented sensitivity and infrared observational capabilities have opened a new window into the earliest luminous sources that existed in the Universe – stars, active black holes, and the galaxies they live in, which were completely impossible to observe with previous telescopes. JWST observations have revealed that the early Universe was extraordinarily, unexpectedly rich with activity, with a large number of galaxies formed only a few hundred million years after the Big Bang, surprisingly massive black holes, and huge amounts of star formation. In this talk, I will discuss some of the breakthroughs that JWST has enabled in studying the early Universe, ask some of the new questions these new discoveries have raised, and briefly describe what comes next in understanding the early Universe with the largest telescopes in space.

Dr Peter Pihlman Pedersen

Dr Pedersen is an alumnus of the Sensor CDT, and is currently a postdoctoral researcher at ETH Zürich. His current research is developing instrumentation and software for observatories, particularly for high-precision near-infrared photometry for detecting and characterising new exoplanets.

Exoplanets: Sensing Methods and Challenges from the Ground Up

The search and characterisation of exoplanets is increasingly defined by the precision of our sensing methods and our ability to control environmental, instrumental, and astrophysical noise. While space missions offer exceptional precision, ground-based observatories remain indispensable—and cost-effective—for continuous monitoring and the development of new sensing techniques. This talk presents recent advances in ground-based exoplanet detection, highlighting methods to mitigate atmospheric variability and the design of next-generation near-infrared instrumentation that enhance the sensitivity and reliability of planetary sensing from Earth.

Prof. Louise Hirst

Prof. Hirst is Professor of Material Physics at the Cavendish Laboratory, University of Cambridge. She formed the Space Photovoltaics research group in Cambridge in 2018, and her current research focuses on the development of photovoltaic systems for next-generation space power applications, including satellite networks, space exploration and space-based solar power.

New applications driving innovation in space power systems

Space based services are increasingly critical to everyday operations across a range of sectors, enabling emergency services and disaster relief, transport and navigation, communications networks, agriculture and meteorology, logistics and financial transactions. Satellite payloads delivering space-based services are powered by photovoltaic systems. Emerging applications such as quantum satellite networks and space based solar power are driving hardware innovation, requiring a unique combination of design criteria including extended radiation resilience for longer missions in more hostile environments; high specific power for further launch cost reduction with high power payloads; deployable form factors for efficient launch packing; and lower fabrication costs. This talk will discuss how new demands are driving innovation, particularly the development of ultra-thin geometries with nanophotonic integration offering inherent radiation resilience.

Student talks

Jake Stuchbury-Wass

Ground reaction force estimation from earables

Jake Stuchbury-Wass, Mathias Ciliberto, Kayla-Jade Butkow, Qiang Yang, Yang Liu, Ezio Preatoni, Dong Ma, Cecilia Mascolo

Ground Reaction Forces (GRFs) are the forces exerted between the foot and the ground during movement. They are fundamental indicators of performance, injury risk, and lower-limb loading in both sports and daily activities. Traditionally, GRFs are measured with force plates in controlled laboratory settings, but these systems are expensive and impractical for real-world use. Recent work has explored wrist-worn devices for gait analysis, yet their signals are often contaminated by hand motion. In contrast, modern consumer earables are increasingly equipped with Inertial Measurement Units (IMUs) for head tracking and interaction. Their head-mounted placement, widespread adoption, and typical paired configuration present a promising but underexplored opportunity for GRF estimation. Here I will present work on our methods for estimating GRF and gait parameters using ear-worn devices. This work is validated through extensive experimentation in a gait lab involving over 35 human participants in a number of different scenarios. We developed signal processing algorithms and machine learning approaches to get accuracies comparable to purpose made devices that lack adoption. We have also worked on making the system portable and run in real time on a phone and earbud set of devices.

Aparna Kaaraal-Mohan

Single-cell analysis of the effects of cellular dormancy on the efficacy of bacteriophages

Aparna Kaaraal Mohan, Ruizhe Li, Temur Yunusov, Diana Fusco, Somenath Bakshi

Bacteriophages, viruses that prey on bacteria, are important regulators of microbial ecosystems in nature and have profound potential as biocontrol agents. However, as obligate parasites their effectiveness is strictly dependent on the metabolic activity of their target cells. In nature, bacterial cells often exist in dormant states and in clusters of densely packed cells, which could limit phage access and activity. Here, we use microfluidics and single-cell measurement of phage activity to investigate how the density and dormancy of cells affect the efficacy of phages.

We find that in deep-dormant E. coli cells, the model phage T7 enters the mode of pseudolysogeny where the host is refractory to lysis but not infection. However, upon resuscitation of the host cells, T7 proceeds to replicate and successfully lyse them. By imaging varying cell densities of E. coli using microfluidic chambers with specified loading capacity, we have also found that dormant E. coli exit the stationary phase earlier when present at high cell densities, indicating possible quorum sensing between the cells.

This finding has significant implications in treating antibiotic-persistent bacterial biofilms or wounds through phage-treatment where bacteriophages can be employed in combination with antibiotics to eradicate antibiotic-resistant bacteria and combat hard-to-treat chronic infections.

Juliana Ferraro

Aerosols, Hydrogels and the Lung: Engineering Tools Towards Precision Drug Delivery

Juliana Ferraro, Róisín M. Owens, Ronan Daly

Respiratory diseases are among the leading causes of death worldwide [1]. Aerosolized drug delivery represents a promising therapeutic route; however, improved understanding of drug screening and delivery efficiency is needed. To this end, physiologically-relevant in vitro models are needed to accurately replicate the microenvironment in the lower respiratory tract and pleural space, to study therapeutics for diseases ranging from asthma to mesothelioma.

In this PhD project, I study the clinical procedures required for aerosol delivery of therapeutics by (i) creating new tissue phantom models, (ii) studying the individual microscale droplets as they interact with the tissue phantom and (iii) examining transport of therapeutics to a target tumour.

Agarose and polyacrylamide hydrogels were formulated to replicate the pleura, a membrane lining the lungs. The mechanical, rheological and diffusive properties of these tissue mimicking materials were characterized. Droplet–hydrogel interactions were then visualized using a custom high-speed, high-magnification imaging platform optimized for micron-scale inkjet-generated droplets, revealing the effect of the substrate (composition, thickness) and droplet characteristics (size, velocity) on impact dynamics. These results establish a foundation for advanced hydrogel-based in vitro models incorporating aerosol exposure and sensing to accelerate translation in precision pulmonary drug delivery.

[1] World Health Organization. “The top 10 causes of death”, (2020)

Jieni Wang

Modelling FUS-R495X-Associated ALS using iPSC-Derived Human Motor Neurons to study gain of toxicity mechanisms

Jieni Wang, Gabriele Kaminski Schierle, Ole Paulsen

Truncating mutations in the Fused in Sarcoma (FUS) gene, such as the R495X variant, lead to clinically aggressive early-onset ALS. The precise mechanisms by which mutant FUS leads to motor neuron degeneration are not fully understood, partly due to the limitations of existing in vitro and in vivo models in recapitulating the complex human disease pathology. Our objective is to establish and characterize an induced pluripotent stem cell (iPSC)-derived motor neuron model for FUS-ALS carrying the R495X mutation. We will investigate key pathological features in the FUS-R495X motor neurons compared to isogenic controls including neuronal stress, calcium handing and cellular organelle morphology using fluorescent and super-resolution microscopy. We anticipate that iPSC-derived motor neurons carrying the FUS-R495X mutation will exhibit characteristic ALS phenotypes, including reduced survival and neurite complexity, cytoplasmic mislocalization of FUS protein, compromised calcium handling, and motor neuron vulnerability to stimulation. We also expect to identify specific cellular pathways perturbed by the R495X mutation that contribute to motor neuron vulnerability. This preliminary work aims to validate a physiologically relevant human cellular model for FUS-ALS, allowing detailed assessment of this truncation-specific toxicity in motor neurons, as well as establishing a tool for pharmacological screening of potential therapeutic compounds targeting this aggressive form of ALS.

Eleni Papafilippou

Epithelial Junction Dynamics Under Cyclic Loading: A Stochastic Framework for Rupture and Healing in Epithelial Tissues

Eleni Papafilippou, Alessandra Bonfanti, Guillaume Charras, Alexandre Kabla

Tissue integrity is fundamental to organ health, yet epithelial tissues are continually challenged by dynamic mechanical environments. In epithelial tissues, cohesion is maintained by intercellular adhesion proteins that form dynamic links between cells. This raises two questions:

• how do oscillatory loads influence the mechanical integrity of epithelial tissues?
• to what extent, does the ability of a tissue to recover or fail depend on the emergent mechanisms of intercellular proteins?

MDCK monolayers were cultured on a custom-built stretching device and cycles of tension were implemented using a PID strain-controlled feedback loop. When subjected to constant stress, monolayers rupture after ~1100s on average. In the cyclic loading condition monolayers spent ~1850s on average at high tension (*P<0.05). To gain insight into the molecular origin for this phenomenon, we examine whether emergent bond dynamics could underlie the delayed failure observed under cyclic loading, using a stochastic, multi-bond model based on Bell's theoretical framework of receptor-ligand binding. Our model reveals three regimes of epithelial stability under cyclic loading: (i) stable recovery, (ii) slow damage accumulation and (iii) rupture. The interplay between extrinsic loads and intrinsic molecular kinetics governs the transition between rupture and healing. From a physiological standpoint, this signifies that soft biological tissues can transiently accommodate prolonged periods of overload. 

Posters

Abbie Aleksandrova

Group A

Mitochondrial Targeting of Nitrogen-Vacancy Centred Nanodiamonds for Quantum Sensing Opportunities

Abbie Aleksandrova, Jack Hart, Sophia Belser, Alex Evtushenko, Mete Atature, Ljiljana Fruk, Helena Knowles

Nanodiamonds (NDs) containing nitrogen-vacancy (NV) centres have shown great ability in measuring modalities, such as temperature or spin species, on the nanoscale due to their unique magneto-optical properties [1]. NVs show superiority over cellular dyes due to their robustness to physical and chemical changes and their lack of photobleaching. The amenability of the NDs to surface functionalization is an attractive feature to explore mitochondrial specific targeting. Mitochondria are a vital organelle in the cell and are interesting for sensing purposes due to their metabolic activity. In particular, mitochondria are a primary source of reactive oxygen species (ROS) which, in high levels, can be an indictor of cell stress and can induce cell senescence [2].

This work aims to achieve nanodiamond mitochondrial targeting through surface functionalisation to enable quantum sensing of mitochondrial activity by the properties of the NV centre.

[1] Gu, Qiushi, et al. ""Simultaneous nanorheometry and nanothermometry using intracellular diamond quantum sensors."" ACS nano 17.20 (2023): 20034-20042.
[2] González‐Gualda, Estela, et al. ""A guide to assessing cellular senescence in vitro and in vivo."" The FEBS journal 288.1 (2021): 56-80.

Filip Ayazi

Group B

Refined contour analysis of erythrocytes

Filip Ayazi, Jurij Kotar, Julian Rayner, Pietro Cicuta

The mechanical properties of red blood cell (RBC) membranes are critical to their function in oxygen delivery and changes to these properties have significant health effects. The function of RBCs requires them to be very soft, which together with the small size of the cells brings the energy required for measurable deformation of the membrane into the range of typical thermal energies and the cells are seen to flicker under optical microscopy. By extracting the shape of the cell's equator and obtaining mean power spectrum by averaging the power present in the normal modes of the thermal fluctuations over time, tension and bending modulus of the membrane can be obtained. We address several issues with this technique. The amount of manual labour required is significantly reduced by developing an automated system allowing the imaging of thousands of cells. We develop a method to improve separation between thermal and slower non-thermal changes using high pass filtering and compensation for loss of thermal mode power, validating this approach with simulations, and finally we test several described methods for correcting for the effect of finite camera exposure time and combine these methods to most effectively compensate for this effect.

Francisco Quero Lombardero

Group A

Toward Scalable Artificial Cell Models: Integrating RNA Condensates and Electrowetting-on-Dielectric

Francisco J. Quero

Understanding how biological complexity arises from molecular interactions remains a challenge. This project develops an experimental framework to systematically explore biochemical behavioural spaces and quantify emergent complexity in artificial cellular systems. The approach integrates three technologies: artificial cells implemented as Electrowetting-on-Dielectric droplets, Active-Matrix Digital Microfluidics enabling combinatorial reagent mixing, and Quality-Diversity algorithms guiding parameter-space exploration to map the system’s behavioural landscape. As a pilot model, programmable RNA condensates emulate cellular compartmentalisation, providing a rich behavioural space for testing the framework. Objectives include: (1) validating RNA condensation in AM-DMF droplets, (2) building an automated platform, (3) implementing closed-loop experimentation for large-scale exploration, and (4) developing complexity metrics to quantify emergence. First-year results show successful platform implementation, generating a library of 100 RNA condensate formations clustered into 10 behavioural categories. Future work will expand this library through closed-loop experimentation guided by MAP-Elites and identify minimal models linking microstates to macrostates, using model simplicity as a proxy for system complexity.

Gratsiela Kostova

Operando Optical Probing of Battery Materials

Group B

Gratsiela Kostova, Arvind Pujari, Michael De Volder, Clare Grey

Lithium-ion batteries play an important role in energy storage solutions, which will be utilised for the transition to net zero. Improved capacity, low cost, reduced environmental impact, and a transparent supply chain are key considerations when scaling battery production. Material selection and performance must therefore be tailored to meet these demands. Operando optical probing of battery materials can monitor the behaviour of particles during charge and discharge.[1] Many battery materials display electrochromic behaviour, where the intercalation of charge causes the generation of different electronic absorption bands in the optical spectrum. Diffuse Reflectance Spectroscopy (DRS) is a technique which employs this behaviour to probe the surface of the electrode during operation. In this work, we report the use of an optimised coin cell geometry which allows for a representative battery performance and optical access to the electrode surface furthest from the separator. [2] DRS allows the reconstruction of the electrochemical cycling data of the cell. This includes the monitoring of phase transitions and oxidation changes in the materials. It provides further information, such as insight into electrode diffusion kinetics and the behaviour of constituent materials in blended electrodes during cycling.

[1] A. J. Merryweather, C. Schnedermann, Q. Jacquet, C. P. Grey, and A. Rao, ‘Operando optical tracking of single-particle ion dynamics in batteries’, Nature, vol. 594, no. 7864, pp. 522–528, June 2021, doi: 10.1038/s41586-021-03584-2.

[2] A. Pujari, B. Kim, N. C. Greenham, and M. De Volder, ‘Identifying Current Collectors that Enable Light–Battery Interactions’, Small Methods, p. 2301572, May 2024, doi: 10.1002/smtd.202301572.

Jack Davies

Group A

Mapping mass transfer and chemical composition within working catalytic systems

Jack Davies, Lynn Gladden, Andy Sederman, Mick Mantle, Qingyuan Zheng, Jordan Ward-Williams

The performance of heterogeneous catalyst pellets is governed by the composition and distribution of material within their pore spaces. Consequently, the overall conversion and selectivity of commercial reactors depend on local reaction kinetics and the multiscale mass transport processes that determine the spatial distributions of material at the pellet-scale. Despite this, understanding of pellet-scale behaviour under operando (i.e. in-use) conditions remains limited, partly due to the lack of experimental techniques capable of probing coupled reaction and diffusion at industrially relevant length scales. Magnetic Resonance techniques are uniquely suited to conduct quantitative measurements of spin density, nuclear spin relaxation, mass transport, and conversion under operando conditions. This work aims to achieve a step-change in the attainable spatial resolution of such measurements, targeting sub-100 μm isotropic resolution in 3D. To date, advanced compressed sensing and image reconstruction techniques have been applied to obtain spatially resolved maps of spin density, molecular diffusion and nuclear spin relaxation times within a single catalyst pellet, achieving 62.5 μm isotropic resolution. A corresponding 3D model has been developed to predict the composition of intra-pellet material under trickle bed conditions. Future experiments will focus on composition mapping during styrene hydrogenation, aiding future catalyst optimisation and enabling model validation.

Jasper Ward-Berry

Group B

LiFETIME - Development of a modular open-source battery cycler to enable battery reuse in LMICs

J. N. Ward-Berry, E. Papafilippou, R. D. Petrie, N. Spiesshofer, M. J. Watt

Rechargeable lithium ion batteries are a ubiquitous part of modern life finding use in everything from smartphones to EVs. As a greater number of rechargeable devices reach end-of-life there is an increasing need to handle waste batteries is a safe and productive manner. Currently the majority of batteries are disposed of in landfill, with only a small fraction being recycled or repurposed, despite many batteries having significant useable capacity at time of disposal. This is a particular problem in LMICs which receive a large proportion of global e-waste while lacking the capacity to properly process it.

LiFETIME is seeking to address this with low-cost open-source battery testing technology. Beginning in 2023 as a Sensor CDT Team Challenge LiFETIME has since spent two years iterating upon and developing a modular battery cycling device to allow the testing of lithium-ion cells for reuse in second-life applications. Our latest prototype can cycle cells in constant-current and constant-voltage modes at up to 3A with current and voltage measurement at a resolution of 1mA and 1mV, for a low-volume production cost of <£25.

With the completion of our most recent prototype LiFETIME has been running a co-creation project in collaboration with Kilele Accelerator, a Nairobi based social enterprise working to create sustainable livelihoods for African youth, to test the local production of our technology and investigate how it could be used to help create sustainable businesses.

Josephine Tumwesige

Group A

Direct-write capacitive touchscreen fabrication to provide insights into fringe field interactions and affordable sensing

Josephine Tumwesige, Dushanth Seevaratnam, Elizabeth A.H. Hall , Ronan Daly

We investigate the direct-write fabrication and characterisation of capacitive touchscreen–derived sensors, focusing on the control of fringe electric fields for selective and quantitative sensing. Coplanar interdigitated electrodes and two-layer “Manhattan” and diamond architectures are patterned by functional inkjet printing, aerosol jet printing, and laser ablation. The electrode width, spacing, and multilayer configuration are systematically varied to tune the field distributions. Structural and electrical measurements are coupled with finite-element simulations and benchtop assays to establish structure–field–response relationships.

Optimised geometries exhibit increased sensitivity and reduced response times to target electrolytes relative to baseline touchscreen structures. Using droplet deposition as the transduction interface, limits of detection in the micromolar range are achieved. The platform leverages established touchscreen form factors and materials, permitting low-cost fabrication and direct coupling to on-device computation and data transfer.

These results indicate that controlled fringe-field engineering on touchscreen platforms can support in situ environmental and agricultural sensing (e.g., metal-ion contamination and nutrient monitoring) with quantitative readout, and provide a framework for linking electrode geometry to sensing performance on commercially relevant substrates.

[1] Diming Zhang, Qingjun Liu. Biosensors and Bioelectronics, 75, 273 (2016).

[2] Sebastian Horstmann, Cassi J. Henderson, Elizabeth A.H. Hall, Ronan Daly. Sensors and Actuators B: Chemical, 345, 130318 (2021)

Justas Brazauskas

Group B

Digital Twins: from sensor design to real-time data visualization

Justas Brazauskas

My thesis explores how digital twins can be designed and evaluated from a human-centred perspective. While digital twin research often focuses on systems engineering, this work bridges that with Human–Computer Interaction and usability science. Using the Cambridge University Computer Laboratory’s LT1 lecture theatre as a Smart Building Testbed, I deployed the Adaptive City Platform—a real-time, event-driven framework integrating edge-AI sensing and data management. On this basis, I created 21 visualisations spanning different levels of fidelity, timeliness, and aggregation. Two frameworks—DiTTo (Digital Twin Taxonomy) and PUX (Patterns of User Experience)—were developed to evaluate and classify these visualisations. Five studies involving Facility Managers, Digital Twin Practitioners, and Building Occupants tested these frameworks in both controlled and ethnographic settings. Results show that task demands, rather than user role, shaped visualisation preferences: 2D views suited high-timeliness expert tasks, while 3D, high-fidelity representations supported communication. Event-driven architectures and composite visualisations improved situational awareness, while privacy-preserving aggregation promoted inclusivity. Together, these findings establish a replicable model for designing adaptive, privacy-conscious, and user-centred digital twins.

Leon Brindley

Group A

Field-Induced Modulation of Conductivity and Seebeck Coefficient in Organic Electrochemical Transistors

Leon Brindley, Dionisius Tjhe, Ian Jacobs, Xinglong Ren, Mohamad Sedghi, Henning Sirringhaus

Organic semiconductors provide a highly versatile platform for next-generation sensors, combining mechanical flexibility, solution processability, and tunable electronic properties. Doping typically enhances their electrical conductivity (σ) by increasing the density of states (DoS) near the Fermi level, facilitating more efficient charge transport. However, this is often accompanied by a reduction in the Seebeck coefficient (S), as the average entropy per carrier decreases at higher carrier concentrations, leading to an intrinsic performance trade-off. In π-conjugated polymers such as IDT-BT, Coulomb interactions between localised carriers can induce a soft gap in the DoS, significantly influencing both σ and S.

We investigate this interplay using dual-gated organic electrochemical transistors (OECTs), building upon architectures widely employed in sensors for their low operating voltages and facile functionalisation. At moderate temperatures, the Fermi level remains pinned to the centre of the Coulomb gap, whereas at low temperatures, field-effect doping can shift it towards the edges. This not only increases σ by moving carriers away from the local minimum in the DoS, but concurrently enhances S by increasing the slope of the DoS near the Fermi level. By jointly fitting σ and S within Fritzsche’s framework of thermoelectric transport, we approximate the shape of the DoS and propose novel strategies to optimise both thermoelectric and bioelectronic devices.

Livia Occhipinti

Group B

Multispecific Antibody Stoichiometry at the Single-Molecule Level

Livia J. Occhipinti, Joseph S. Beckwith, Ezra Bruggeman, Brendan Cullinane, Praveen Kallamvalli Illam Sankaran, Steven F. Lee

Multispecific antibodies (msAbs) are next-generation immunotherapeutics designed to simultaneously target multiple antigens, offering powerful new strategies in oncology. While their sophisticated structures are crucial for tackling cancer heterogeneity, they also introduce significant development challenges, such as the generation of mispaired variants that compromise the final yield and therapeutic quality[1,2]. While conventional bulk approaches average over a population of molecules, single-molecule fluorescence microscopy (SMFM) offers the sensitivity required to resolve molecular heterogeneity and capture the full distribution of molecular states[1-4].

In this work, we present an optimised SMFM workflow that combines site-specific fluorescent labelling with advanced image analysis to precisely characterise individual antibodies. The stoichiometry is determined using multiple tools, such as stepwise photobleaching, blinking analysis, and photon count distributions. A digital twin framework is also built to simulate entire single-molecule experiments by incorporating photophysical properties and imaging parameters, and enabling the prediction and optimisation of experimental conditions prior to acquisition[3,4].

These developments establish a refined methodology for stoichiometric studies of monoclonal antibodies, and lay the crucial foundation for extending it to msAbs, and integrating it as a real-time, high-throughput quality control sensor tool during manufacturing.

[1] Elshiaty, M.; Schindler, H.; Christopoulos, P. Principles and Current Clinical Land scape of Multispecific Antibodies against Cancer. International journal of molecular sciences 2021, 22, 5632.

[2] Wei, J.; Yang, Y.; Wang, G.; Liu, M. Current landscape and future directions of bispecific antibodies in cancer immunotherapy. Frontiers in immunology 2022, 13, 1035276.

[3] Lelek, M.; Gyparaki, M. T.; Beliu, G.; Schueder, F.; Griffi´e, J.; Manley, S.; Jung mann, R.; Sauer, M.; Lakadamyali, M.; Zimmer, C. Single-molecule localization microscopy. Nature Reviews Methods Primers 2021, 1, 39.

[4] McLean, A.; Sala, R.; Longbottom, B.; Carr, A.; McCune, J.; Lee, S.; Scherman, O. Single-Molecule Stoichiometry of Supramolecular Complexes. Journal of the American Chemical Society 2024, 146, 12877–12882.

Max Walk

Group A

Investigating RNA-RNA interactions in rotavirus genomic packaging

Aidan Tollervey & Max Walk, Michael Knight, Emilia Argüello, Lee Sherry, Alexander Borodavka

Rotaviruses are pathogenic viruses that possess segmented genomes made up of eleven double-stranded RNA segments, each selectively packaged into individual virions. The exact mechanism of this assortment is unclear, however, it is becoming increasingly apparent that RNA-RNA interactions are involved in driving this assortment.

To determine which interactions may be involved in the assortment of the rotavirus genome, we employed RNA structure probing techniques. We used the sequencing of psoralen crosslinked, ligated, and selected hybrids (SPLASH) method to directly observe RNA–RNA interactions among rotavirus transcripts. This approach involves treating RNA with psoralen, which, upon UV irradiation, induces crosslinking, and subsequently sequencing the resulting hybrid RNA to map both intra- and intermolecular interactions. To analyse the SPLASH data, a bioinformatics pipeline was developed using the Nextflow workflow management software. The pipeline automates several steps, including data pre-processing, read alignment, and identification of putative RNA-RNA duplexes.

This ongoing work is focused on interrogating key differences in the formation of short-range and long-range RNA–RNA interactions, as well as the role of co-transcriptional RNA folding in the mechanism of genomic RNA assortment.

Melissa Watt

Group B

Examining the potential of short-wave infrared wavelengths to minimise bias in wearable health monitoring

Melissa J. Watt, Layla Malouf, Isabelle Racicot, Ran Tao, Janek Gröhl, Sarah E. Bohndiek

Wearable devices such as smart watches provide monitoring of vital signs and enable pulse oximetry using light emitting diodes in the visible and near-infrared (NIR). A key challenge with the current generation of wearables is that they are only sensitive to surface vasculature and are susceptible to measurement bias due to melanin absorption in the skin. Due to lower scattering, short-wave infrared (SWIR) wavelengths have been proposed to penetrate deeper into biological tissue, and could avoid confounding skin tone bias, but have yet to be used in commercial wearables.

Here, we undertook a detailed analysis of the SWIR absorption and scattering properties of common tissue constituents: whole blood, haemoglobin, lipids, melanin and water, using single integrating sphere systems. Using these results, we created layered tissue models with diverse skin types in silico probed with Monte Carlo simulations which showed reduced penetration depth for darker skin tones at visible wavelengths, whilst at SWIR wavelengths, penetration depth was similar regardless of skin tone. To validate simulations and enable robust testing of wearable prototypes, we fabricated tissue-mimicking, pigmented skin layer phantoms (~ 1 mm thick) using a co-polymer-in-oil material formulation with tunable optical and acoustic properties. Digital twins of these phantoms were used to show agreement with the layered tissue model simulations.

[1] Assorted Spectra compiled by Prahl & Jacques, available at https://omlc.org/spectra/index.html

[2] Gröhl J et al., 2022 J. Biomed. Opt. 27 083010

[3] Else T R et al., 2023 J. Biomed. Opt. 29, S11506.

Oriol Colomer I Ferrer

Group A

Nanoelectrum: Ag-Au alloy nanoparticles with controlled size, shape and composition

Oriol Colomer i Ferrer, Matthew Ellis

The poster presents nanoelectrum (Ag–Au nanoalloys), which combine Ag’s field enhancement with Au’s stability. Adoption is limited by current syntheses that cannot independently tune size, shape, and composition or ensure long-term structural stability. We introduce two room-temperature, seed-mediated routes that span a wide Au composition range and yield a single, composition-dependent LSPR tunable across the visible spectrum. A surfactant route (CTAC) and a surfactant-free route (NaCl) reveal the key levers linking synthesis to performance: the capping agent, the strength of the reducing agent, and post-synthetic heating. We also demonstrate that glutathione (GSH) stabilises hollow Ag–Au nanoparticles in the long term. In comparison, CTAC delivers greater colloidal stability at higher Au content and with larger seeds, whereas the NaCl route is a robust surfactant-free alternative for settings where surfactants are undesirable. This study elucidates practical design rules for Ag–Au nanoparticles (Ag–AuNPs) with tunable optical response and long-term stability, supporting applications in catalysis, biosensing, biomedicine, and SERS detection.

Panagiotis Ioannou

Group B

HyperGenie: A method for predicting enzymatic gene essentiality using hypergraph neural networks and genome-scale metabolic models

Panayiotis Ioannou, Iulia Duta, Suraj Verma, Pietro Cicuta, Pietro Lio, Claudio Angione

We introduce HyperGenie, a method for predicting enzymatic gene essentiality by leveraging hypergraph neural networks and genome-scale metabolic models. In this approach, the E. Coli metabolic network is represented as a hypergraph, with metabolites as nodes and reactions as hyperedges. HyperGenie achieves significant performance gains over existing methods, improving the area under the precision-recall curve (PR-AUC) by over 12% for E. Coli growing on glucose under aerobic conditions. It matches or surpasses the accuracy of flux balance analysis without assuming that knockout strains optimise for growth. Notably, HyperGenie maintains a PR-AUC above 90% using only 40% of the training data and exhibits robust performance across various nutrient conditions.

Raluca-Elena Alexii

Group A

Investigating Protein-RNA Interactions by Solid-State Nanopore Sensing

Raluca-Elena Alexii, Catherine Wilson and Ulrich Keyser

mRNA as a therapy has gained huge traction in the past decade, with many efforts being made to optimise exogenous transcripts for translation in-vivo. This includes the incorporation of synthetically designed UTRs, modified nucleotides and, extensive codon optimisation. However, RNA binding proteins (RBPs) also play profound roles in the mRNA lifecycle. The extensive responsibilities of RBPs range from intranuclear processing and modification to nuclear export and subsequent stability, localisation and translation. Despite the significance of mRNA proteomes, they have not yet been taken into consideration in the design and optimisation of mRNA therapeutics. This project considers the proteomes of endogenous, exogenous and codon optimised mRNAs through the development of a novel solid-state nanopore sensing methodology. Our technology detects the novel phenomena and binding stoichiometries which remain obscured by current ensemble methods. Through the development of a controlled 1:1 stoichiometric protein:RNA interaction we will show that solid-state nanopores can be used to successfully characterise such complexes. Furthermore, optimisation of the system’s operational parameters will increase binding site resolution whilst targeted enzymatic degradation of background species will allow for the detection of low abundance species from complex mixtures. Overall, this project aims to add an additional methodology to the growing repertoire of solid-state nanopores.

Robert Petrie

Group B

Flexible Agarose Gel Waveguides for Oceanographic pH Sensing

Robert Petrie, Ishtiaq Ahmed, Tijmen Euser, Ljiljana Fruk, Oscar Branson

Developing simple, robust, non-toxic, and reusable pH sensors is of critical importance to ocean acidity monitoring. We designed a flexible, biocompatible waveguide made of agarose gel and functionalised it with the standard oceanographic colourimetric pH indicator, meta-Cresol purple, providing an integrated sensing solution.

Traditional silica waveguides are performant and reliable bases for many sending applications. However, the creation of fibre-optic fluid sensors typically involves complex post-processing, such as tapering, cladding replacement, or the use of Bragg gratings. Further, silica suffers from limited biocompatibility, rigidity, and complex functionalisation methods. These limitations often mean that fibres are used only for optical coupling with the sensor, rather than forming part of the sensor itself. Methods to achieve this coupling can often be complex and limit the system’s effectiveness.

In contrast, agarose gel is biocompatible and exhibits characteristics amenable to optical sensing – transparency, a tuneable refractive index, flexibility, and many functionalisation pathways. To demonstrate the possibilities of gel sensors, we have developed flexible agarose-based waveguides, functionalised with meta-Cresol Purple, to produce an optically integrated, colour-changing pH sensor. The fluid sample (seawater) infiltrates the open structure of the gel and interacts directly with the indicator dye, enabling waveguide-enhanced spectroscopy.

Sofia Kapsiani

Group A

Deep learning for fluorescence lifetime predictions 

Sofia Kapsiani, Nino F. Läubli, Edward N. Ward, Ana Fernandez-Villegas, Bismoy Mazumder, Clemens F. Kaminski, Gabriele S. Kaminski Schierle

Fluorescence lifetime imaging microscopy (FLIM) is a powerful microscopy technique for studying biological systems at a molecular level. FLIM is widely used across various fields, including cancer research, neurodegeneration, and plant science. However, data collection and interpretation are often challenging, and traditional data analysis methods require a high number of photons per pixel for reliable fluorescence lifetime measurement. This results in prolonged data acquisition times, making FLIM a low-throughput technique with limited capability for in vivo applications. Here, we introduce a deep learning model capable of quantifying FLIM data obtained from photon-starved environments[1]. Our method outperforms other published deep learning approaches[2,3] and traditional data analysis techniques, accurately calculating fluorescence lifetimes from decay curves with fewer than 50 photons per pixel and thus, shortening FLIM acquisition times to just a few seconds. Following characterisation on both simulated data and experimental data, and we have demonstrated its ability to analyse data from live, dynamic samples. As a case study, we quantified disease‑related protein aggregates in live Caenorhabditis elegans, extending FLIM applicability to longitudinal studies without the need of chemical immobilisation that can disrupt biological processes.

[1] Kapsiani, S., Läubli, N.F., Ward, E.N., Fernandez-Villegas, A., Mazumder, B., Kaminski, C.F. and Kaminski Schierle, G.S., 2025. Deep learning for fluorescence lifetime predictions enables high-throughput in vivo imaging. Journal of the American Chemical Society.
[2] Smith, J.T., Yao, R., Sinsuebphon, N., Rudkouskaya, A., Un, N., Mazurkiewicz, J., Barroso, M., Yan, P. and Intes, X., 2019. Fast fit-free analysis of fluorescence lifetime imaging via deep learning. Proceedings of the national academy of sciences, 116(48), pp.24019-24030.
[3] Xiao, D., Sapermsap, N., Chen, Y. and Li, D.D.U., 2023. Deep learning enhanced fast fluorescence lifetime imaging with a few photons. Optica, 10(7), pp.944-951.

Stephen Devlin

Group B

An Integrated Platform for Automated, High‑Throughput Soft‑Matter Experiments

Stephen Devlin

Soft-matter and bottom-up synthetic-cell studies are hampered by large design spaces and manual experimentation. This automated platform integrates robotic liquid handling, a custom multi-well plate with well-addressable temperature control, and fluorescence microscopy. A Python controller coordinates dispensing, thermal profiles, and imaging while persisting data in a relational model with end-to-end provenance. Using open standards and a well-factored class model, the platform is easily extensible, providing a practical template for self-driving labs in biomolecular engineering.

Nicolas Spiesshofer

Group A

Tailoring giant-index plasmonic metamaterials for high-sensitivity surface-enhanced IR and Raman spectroscopies

Nicolas Spiesshofer, Elle Wyatt, Zoltan Sztranyovszky, Caleb Todd, Taras Mykytiuk, James W. Beattie, Rowena Davies, Rakesh Arul, Viv Lindo, Thomas F Krauss, Angela Demetriadou, Jeremy J Baumberg

Mid-infrared (MIR) spectroscopy has been revolutionised by the development of plasmonic metamaterials, allowing enhanced selective molecular sensing in situ, label-free and at low concentrations [1]. However, achieving easily assembled materials with both high enhancement factors and well-controlled geometries remains a significant challenge. Multilayer aggregates (MLaggs) of close-packed gold nanoparticles (AuNPs) exhibit ultrastrong extinction resonances that enable both surface-enhanced IR absorption (SEIRA) and Raman (SERS) spectroscopies, providing complementary sensing capabilities [2,3]. Here, we show through optical simulations and experimental validation that these readily fabricated MLaggs behave as low-loss, high-refractive-index (n) metamaterials in the MIR [4]. Critical geometrical parameters, including interparticle gap size, layering and faceting, are key factors influencing optical behaviour, leading to ultrahigh refractive indices exceeding n > 10 without excessive loss, a combination not found in natural materials, and expands electromagnetic interactions in these structures in unprecedented ways.

We further create tunable combinatorial metamaterials with intense light-matter coupling by co-aggregating silver (AgNPs) and gold nanoparticles (AuNPs). Controlled chemical dissolution of variable concentrations of AgNPs in these structures enables tuning of the fill fraction, resonance wavelength, and refractive index, revealing a direct relationship between structural porosity and optical behaviour. This tunability is consistent with our MLagg metamaterial model and enables control over SEIRA enhancements.

Increasing porosity strongly improves the detection of surface-bound analytes with SEIRA spectroscopy and SERS, demonstrated here using low concentrations (2.5 mgml^-1) of 50 nm nanoparticles of polystyrene, an environmentally abundant synthetic plastic. These results encapsulate the design trade-off between the MIR resonance position and analyte accessibility, which is optimised here for sensing in the molecular fingerprint region. These metamaterials thus show great potential for a wide range of applications in healthcare, chemical sensing, and pollutant monitoring.

[1] A John-Herpin et al. Adv. Mater. 35, 2110163 (2023)
[2] R Arul et al. Light Sci Appl 11, 281 (2022)
[3] SM Sibug-Torres et al. Nat Comm 15, 2022 (2024)
[4] N Spiesshofer et al. Optica, Vol.12, Issue 9, pp. 1357-1366 (2025)