The 2 Warwick Postdoctoral Science Symposium Book of Abstracts

The 2nd Warwick Postdoctoral Science Symposium Book of Abstracts

Medical Teaching Centre, Gibbet Hill Campus, University of Warwick

7th May 2019 ii Contents

Contributed talks 1 Extreme environments for electrons - The physics of low temperatures and high magnetic ﬁelds (Kathrin Gotze¨ )...... 2 Analyzing the brainstem circuits for respiratory chemosensitivity in freely moving mice (Amol Bhandare )...... 3 Thermal management at nano-scales: modelling of multi-scale and multi- phase ﬂows (Anirudh Rana )...... 4 Increasing yields in biosynthetic pathways using orthogonal ribosomes (Alexander Darlington )...... 5 DGAT1 matters for VLDL size: Implications for coronary artery disease ( Zehra Irshad )...... 6 Quantifying the structure of the mental lexicon using network science ( Cynthia Siew )...... 7 Modelling stress-affected localised chemical reactions without remeshing ( Mikhail Poluektov )...... 8 From Leaf to Skin: a Journey Towards Better Sunscreens ( Michael Horbury )...... 9

Flash talks 11 Robust silver electrodes for highly efﬁcient and stable organic solar cells ( Jaemin Lee )...... 12 Indexing massive datasets ( Dominik Kempa ) ...... 13 How did the tongue evolve in tetrapods? ( Ritika Ghosal ) ...... 14 Prediction of large format high power cell performance using NTG model ( Manikandan Balasundaram )...... 15 Study of protein-protein interactions within the bacterial cell division wall, as a new target for antimicrobial agents ( Ingrid Edwards )...... 16 MicroRNAs in brown and white adipocytes ( Federica Dimitri )...... 17

iii How does the Arabidopsis auxinome respond to auxins applied as herbicides? ( Martin Kubes )...... 18 Aging Characteristics of Lithium-ion Batteries ( Selcuk Atalay ) ...... 19 A machine learning-developed classiﬁcation system of GP practices, and its relationship to antibiotic prescribing. ( Saran Shantikumar )...... 20 Analysing computing performance through modelling and proxy applications ( James Davis ) ...... 21 The potential of an ancient remedy for the treatment of diabetic foot infections ( Blessing Anonye ) ...... 22 In situ Operando Characterisation of Lithium Ion Battery Degradation ( Anisha Patel )...... 23

Posters 25 P01: High resolution imaging of conjugated polymers using combined electrospray deposition and scanning tunnelling microscopy ( Luis Alves Perdigao )...... 26 P02: First-Principles Molecular Dynamics Study of Rechargeable Lithium Based Batteries (Shalini Awasthi )...... 27 P03: Surveying atoms inside nanometer scale wires ( Aruni Fonseka )...... 28 P04: Uterine natural killer cells are regulated by vascular adhesion protein- 1 in the endometrium (Seley Gharanei )...... 29 P05: Mathematical modelling of dilute granular gases (Vinay Kumar Gupta ) ...... 30 P06: Can we end the vicious cycle of antibiotic resistance in CF? ( Marwa Hassan )...... 31 P07: Hidden Markov models for monitoring circadian rhythmicity in telemetric activity data (Qi Huang )...... 32 P08: Macrocyclisation of Small Peptides Enabled by Oxetane Incorporation (Stefan Roesner ) ...... 33 P09: Phospho-regulation of tropomyosin is crucial for actin cable turnover and division site placement ( Saravanan Palani )...... 34 P10: Accumulation of ferric iron in ageing Drosophila brain is independent of macroautophagy ( Anne-Claire Jacomin )...... 35 P11: Mechanism of biological effect of the new photosensitive organometallic iridium(III) complexes ( Jitka Pracharova )...... 36

iv P12: Tracking kinetochores in lattice light sheet microscopy data ( Jonathan Harrison )...... 37 P13: Aberrant Specification Of Endometrial Stromal Cells Into Distinct Decidual Subpopulations In Recurrent Pregnancy Loss. ( Emma Lucas )...... 38 P14: Control of sp2 content in boron doped diamond for quantitative detection of hypochlorite at high concentrations ( Anthony Lucio )...... 39 P15: Chasing gravitational wave sources in the electromagnetic domain ( Joseph Lyman ) ...... 40 P16: Analyses on Representations of Materials for Machine Learning ( Berk Onat ) ...... 41 P17: Modelling rare-earth/transition-metal magnets at finite temperature with self-interaction-corrected relativistic density-functional theory and disordered local moments ( Christopher Patrick ) ...... 42 P18: Symmetry design of Magnetoelectric coupling in Perovskites (Fernando Pomiro )...... 43 P19: Characterisation of the Circadian Modulation of Blood-Brain-Barrier Permeability in an In Vitro Tri-Culture Mode ( Swati Kumar )...... 44 P20: A Boron Doped Diamond Multi-Tool Electrode For Simultaneous pH and Dissolved Oxygen Analysis. ( Tania Read )...... 45 P21: Sharp bound on the number of maximal sum-free subsets of integers ( Maryam Sharifzadeh ) ...... 46 P22: Regaining the ashes: finding chemical ‘clues’ to mitigate the impact of ash dieback ( John Sidda )...... 47 P23: Dose enhancement studies in synchrotron radiation therapy ( Jenny Spiga ) ...... 48 P24: The Warwick Centre for Ultrafast Spectroscopy: Dynamics, the Full Spectrum (Michael Staniforth )...... 49 P25: Bacteriophage: Potentials for enhancing antibacterial therapy ( Eleanor Townsend )...... 50 P26: Transparent metal electrodes for organic solar cells using an efficient lift-off process ( Silvia Varagnolo )...... 51

Author Index 53

v vi Contributed talks

1 Extreme environments for electrons - The physics of low temperatures and high magnetic ﬁelds Kathrin Gotze¨ Superconductivity and Magnetism Group Department of Physics

Superconductivity, quantum mechanical effects and certain types of magnetism can only be observed at very low temperatures. In order to find out more about the physics behind these phenomena, we can manipulate materials that show them by applying high magnetic fields and high pressures. Changes in physical quan- tities such as electrical resistance or magnetisation indicate the response of the electrons of the material and allow researchers to draw conclusions about fundamental mechanisms. But what do physicists mean when they talk about low temperatures? How cold does it get in the lab? And how do you generate a high magnetic field? Can a material change its properties when you compress it? I will introduce the audience to the extreme conditions that we can create in our labs and why we need them to investigate the physics of superconductivity and magnetism.

2 Analyzing the brainstem circuits for respiratory chemosensitivity in freely moving mice Amol Bhandare Dale Group School of Life Sciences

Neuroscience seeks to establish the causal links between neuronal activity and behaviour. To achieve this, an essential step is to document neuronal activity at the single cell level during behaviour. For the analysis of the neural mechanisms of breathing this has been surprisingly difﬁcult: neuronal activity patterns during respiration are mostly investigated in isolated neonatal tissue, or in anaesthetized preparations. Both of these approaches have signiﬁcant drawbacks: the neural cir- cuitry for the control of breathing undergoes considerable maturation postnatally; and anaesthesia, by its very nature, alters the operation of neural circuits and has well documented depressive effects on breathing. We have now achieved some- thing held by many to be impossible - imaging the activity of single neurons in the medulla oblongata of awake unrestrained mice. Application of modern endoscopic methods, combined with genetic targeting of Ca2+ reporters and head-mounted mini-microscopes, has allowed unbiased and simultaneous imaging of the activity of multiple genetically targeted neurons, in the brainstem. We have used our methods to study the chemosensory control of breathing and distinguished the relative importance of two competing chemosensory nuclei -the retrotrapezoid nucleus (RTN) and medullary raphe (bioRxiv doi.org/10.1101/492041). The existing evidence supports the involvement of both nuclei. Surprisingly our data unambiguously show that in awake freely behaving mice, the neurons of the ros- tral medullary raphe are predominant in coding inhaled CO2. By contrast the neurons of the RTN have heterogenous activity patterns in response to inspired CO2, and make a relatively small contribution to the response to hypercapnia.

3 Thermal management at nano-scales: modelling of multi-scale and multi-phase ﬂows Anirudh Rana Micro-Nano Flows Warwick Mathematics Institute

Micro/nano-scale phase transitions are of great importance to many practical applications of science and engineering, for example, thermal management in high power density electronics, energy harvesting, as well as of fundamental interest due to the new physics appearing at small scales. One of the main challenges in the description of micro/nano-scale phase-transition processes is that the classical NSF paradigm fails to accurately capture the vapour flow due to rarefaction effect- sas the length scale of the flow becomes comparable to the mean free path in the vapour. The objective for my talk is to investigate hierarchical extended fluid dynamics models to gain a better understanding of phase-change process at micro/ nano scales.

4 Increasing yields in biosynthetic pathways using orthogonal ribosomes Alexander Darlington Bates Group School of Engineering

The introduction of new functions, encoded by non-host genes, into microbes is a crucial component of synthetic biology and biotechnology. However the hosts cellular resources, such as RNA polymerases and ribosomes, are finite so the introduction of foreign genes creates competition for gene expression machineries. This can lead to the emergence of unwanted coupling between the expression of different genes, which complicates circuit or pathway design and potentially leads to failure. This can significantly increase costs and design timeframes due to iter- ative experimentation and redesign. Previous studies have highlight translational limitations as a key cause of gene coupling. Expression of an engineered 16S rRNA with altered specificity allows translational resources to be partitioned into host specific and orthogonal (circuit-specific) ribosomes. Using a combination of mathematical modelling and experimental implementation we have demonstrated that this separation of activities can be used to alleviate the effects of resource competition and thereby increase the ability of circuit modules to function independently. We have designed and experimentally implemented a dynamic genetic controller, which senses the demand for translational resources by synthetic genes and acts to increase orthogonal ribosome production as demand increases. Our prototype controller significantly reduces resource-mediated coupling in a simple genetic circuit. Our results highlight the potential for dynamic resource allocation as a means of minimising the impact of cellular limitations for improved functioning of gene circuits or metabolic pathways.

5 DGAT1 matters for VLDL size: Implications for coronary artery disease Zehra Irshad Zammit Group WMS, Translational & Experimental Med, Biomed Sci

Increased Hepatic Triglycerides (TAG) synthesis and secretion are associated with hepatic steatosis and dyslipidaemia, two markers of the cardio metabolic syn- drome and Type-2 diabetes. Two main enzymes (diacylglycerol acyltransferases DGATs) that catalyse the final reaction of TAG synthesis have been identified, DGAT1 and DGAT2. They play partly non-redundant roles in different tissues, as highlighted by the very different phenotypes of Dgat1-/- mice, which are resistant to obesity, and Dgat2-/- mice, which die within days of being born. The aim of this study was to ascertain whether DGAT1, which has a different membrane distribution from DGAT2, and is itself located on either side of the endoplasmic reticular (ER) membrane, plays a distinctive role in determining the triglyceride content of very-low-density lipoprotein (VLDL/fat) particles secreted by the liver. DGAT1-LKO mice had the same rate of secretion of VLDL particle, but these particles were approximately half the size of VLDL particles secreted by controls, and had a similarly decreased content of TAG, but with normal cholesterol and cholesterol ester content. The livers of DGAT1-LKO mice showed a specific compensatory increase in the expression of the mRNA for glycerol phosphate acyltransferase 1 (GPAT 1). The effects of DGAT1- and DGAT2-specific inhibitors on apoB secretion by HepG2 cells showed that there is a high, but not complete, redundancy in the ability of the two enzymes to support apoB secretion suggesting that while DGAT1 determines particle size, DGAT2 can fully support particle number. DGAT1 is central to the complete (second-step) lipidation and maturation of VLDL particles within the lumen of the ER. The present findings may provide the basis for dietary recommendations to facilitate the decrease of VLDL size, and its propen- sity to give rise to dangerous levels of particles containing the deleterious type of cholesterol.

6 Quantifying the structure of the mental lexicon using network science Cynthia Siew Department of Psychology

Cognitive representations are central to theoretical and computational accounts of behavioral and language phenomena in many areas of cognitive psychology and psycholinguistics. In this talk I will demonstrate how the tools of network science can be used to quantify and represent cognitive representations as networks of phonological, orthographic, and semantic relationships. I will present results from psycholinguistic and computational experiments, as well as large-scale analyses of megastudies and linguistic norms, that demonstrate (i) how language-related processes are inﬂuenced by the network structure of the mental lexicon, and (ii) how the development of the phonological lexicon occurs over the lifespan.

7 Modelling stress-affected localised chemical reactions without remeshing Mikhail Poluektov IINM Group WMG

The analysis of the stability and the kinetics of phase boundaries in solid-solid phase transitions has a long history in solid mechanics. For materials with complex rheology undergoing large deformations, such analysis must be performed computationally, which is a challenging task. When the standard finite-element method (FEM) is applied to problems with moving interfaces, the geometry should be remeshed each time the interface moves. This technique requires using an extremely fine mesh for achieving a sufficient accuracy and additional computational resources for performing remeshing. Furthermore, the automation of the remeshing process can be a non-trivial task. An alternative way of treating such problems is a computational method that allows the interface to cut through the elements and to move independently of the mesh, the so-called CutFEM approach. Up to now, CutFEM has been formulated for linear problems only. In this talk, a gener- alisation of CutFEM to large deformations and arbitrary constitutive behaviour of materials is presented. It is demonstrated that the proposed implementation of the method has the same convergence rate with respect to the mesh size as the standard FEM. The method is used to simulate the kinetics of phase boundaries and chemical reaction fronts in hyperelastic bodies undergoing large transformation strains.

8 From Leaf to Skin: a Journey Towards Better Sunscreens Michael Horbury Stavros Group Department of Chemistry

The requirement for superior sunscreening agents has become more pertinent in recent years, with the increase ease of being able to holiday in sunnier climates and the significant rise in skin cancer incidents worldwide. This is further com- pounded by several controversies around current sunscreening agents, such as their toxicity and environmental impact. One avenue for developing superior sunscreening agents is to look to how nature deals with excessive ultraviolet exposure; particularly in plants. Plant leaves utilize the sinapate ester sinapoyl malate, to protect them from the harmful effects of ultraviolet, thus, gaining an understanding into how this sinapate ester achieves its intrinsic photoprotection will help in the design of new sunscreening agents. To gain an understanding of how sinapoyl malate works, we have utilized a combination of time-resolved and steady state spectroscopy to probe what occurs after sinapoyl malate absorbs ultraviolet light. However, as it isnt feasible to study it directly in the leaf. Therefore, we have had to use a bottom-up approach to first understand how it behaves in the sim- plest model environment, such as neat solvents. Once the understanding has been gained in these simple models, we can move to models of increasing complexity, ultimately aiming to reach a true to life environment. Using the information, we have gleaned from these studies we have unravelled the mechanism behind sinapoyl malate and sinapate esters high photostability, along with several drawbacks that need to be addressed before they can be used in sunscreen formulas. Recently, we have identified a potential molecule based off these sinapate esters that will hopefully address these concerns. Therefore, we have carried out time- resolved spectroscopy measurements on this new molecule, including, for the first time, using a combination of a commercial sunscreen emollient and a synthetic skin mimic.

9 10 Flash talks

11 Robust silver electrodes for highly efﬁcient and stable organic solar cells Jaemin Lee Hatton Group Department of Chemistry

Silver is a very attractive material for various opto-electronics such as organic photovoltaics (OPVs), organic light-emitting diodes (OLEDs) based on its low ex- tinction coefficient and flexibility as an electrodes on plastic substrates. However, due to the high surface energy of silver and its slow oxidation in air, optically thin evaporated Ag films (<20 nm) are morphologically unstable even at room temperature. Consequently, for practical applications an easily implementable and versatile method of imparting long term stability is required. This talk will demonstrate how a single layer of a small molecule deposited from the vapour phase can greatly enhance the morphological and chemical stability of very thin Ag film electrodes. Due to its very low thickness (∼1 nm) this monolayer does not electrically isolate the electrode. It is shown that in 10% efficient inverted OPVs and semi- transparent OPVs substantial improvements in device stability can be achieved by inclusion of this layer, which is remarkable given its very low thickness.

12 Indexing massive datasets Dominik Kempa Theory and Foundations Group Department of Computer Science

One of the biggest challenges of modern computing is efﬁciently storing and searching massive collections of data. Sources of such data include multi-author databases of documents, such as Wikipedia, collections of versioned source code such as GitHub, and web crawls. More recently, with the advent of high-throughput DNA sequencing machines, enormous volumes of genomic data that requires indexing have been released. A plan of sequencing genomes of up to one hundred thousand citizens was recently announced by the British government. This project will produce half a petabyte of data. In this talk, I would like to give an overview of the state-of-the-art techniques used to store massive datasets. This so-called ”full-text indexing” combines aspects of information theory, data compression, and combinatorial pattern matching aims to structure the data so that it requires small space (close to the data in compressed form) and simultaneously support various types of queries over the underlying data (searching, for example) without decom- pressing the data. The area has been witness to intense research in the past two decades, and several of its fruits are now widely used in molecular biology.

13 How did the tongue evolve in tetrapods? Ritika Ghosal Koentges Lab School of Life Sciences

The tongue is an essential organ present in most tetrapods (limbed vertebrates) but absent in all fish species (limbless vertebrates). The molecular and cellular ori- gins of the tongue muscle patterns at the fish-tetrapod transition have remained elusive to date. A class of unique, embryonic stem cell population called the neural crest cells are thought to be the organizers of these patterns. As these populations are shared across all vertebrates, an experimental inter-species ontogenetic com- parison gives us the unprecedented opportunity to capture the underlying mechanisms for tongue evolution. We performed a detailed comparative migratory mapping of specific groups of neural crest cells at single cell resolution in zebrafish (representing the fish condition) and several tetrapods (urodeles, mice and birds), using transplantation and genetic fate mapping approaches. We then carefully superimposed the cellular/population architectures onto the complex muscular patterns across these vertebrate classes. We traced the common ancient (fish) ori- gins of muscles in the mouse and other tetrapods at the cellular level, revealing an ancient history of cellular patterns 400 MYA. This revealed cryptic cell population compartments hitherto considered ’homogeneous’ that render transitions and muscle fusions visible as the underlying causes of the evolutionary morphological novelty of the tongue.

14 Prediction of large format high power cell performance using NTG model Manikandan Balasundaram Energy Systems Group WMG

Accurate prediction of cell performance using electrochemical parameters has al- ways been a challenge in terms of computational complexity and collection of parameters for individual cell components (such as cathode, anode and electrolyte). In construction of a completely physics based model, the parameterisation is te- dious and extremely time consuming. In order to save time and maintain accuracy, an alternate mathematical model which demands less experimental inputs especially from full cell alone is of interest in this study. In addition, the model which offers scope to manipulate the geometrical features such as tab positions and cell dimensions to further optimise the cell performance at lower computational cost is desirable. NTG model is a semi empirical model which predicts the transient behaviour of lithium ion batteries. In this study, two signiﬁcant full cell parameters namely effective conductance and open circuit voltage are determined for a high power cell by stepped pulse experiments over a wide range of operating temperatures between -25◦C and 50◦C. Other parameters such as cell thermal conductivity, speciﬁc heat capacity and density are also determined. Additional heat balance equations are incorporated to simulate the thermal behaviour of cell. The model is constructed using mathematical PDEs in COMSOL and solved across a representative 3D geometry. Interestingly, the electrochemical performance curves and the cell surface temperature from simulations are quite close to that of experiments. The feasibility of potential, current density and temperature distribution in three dimensions in this model give better insights into the underlying battery physics. Although the diffusion of lithium ion is overlooked in this model for simplicity, the model offers the advantage of faster simulations for high power cells where the electrode loading is lesser.

15 Study of protein-protein interactions within the bacterial cell division wall, as a new target for antimicrobial agents Ingrid Edwards Lewandowski Group Department of Chemistry

A state of emergency was declared by the World Health Organization already four years ago to combat the increasing rate of resistance arising in bacteria to all currently available antibiotics on the market. Bacterial membrane and more importantly bacterial cell division is still a domain showing the least resistance mechanisms, making it a valuable target for new antibiotics. Bacterial cell division is orchestrated by a group of membrane bound and cytoplasmic proteins, which form a complex known as the Divisome. At the core of the divisome is the bacterial ho- mologue of tubulin; FtsZ, which polymerises and exerts a cytokinetic force on the cell membrane that leads to cell division. FtsZ contain a GTP-binding domain and in the presence of GTP, FtsZ can assemble into protofilaments, two-dimensional sheets, and ring structures and overall FtsZ and tubulin share substantial structural similarities. The FtsZ-associated proteins ZapA to ZapE interact at an early stage of FtsZ-ring assembly. Previous work has shown that the cell division protein ZapA needs to tetramerise into a dog bone shape in order to achieve ordered assembly of FtsZ fibres. The interaction of ZapA decreases the overall GTPase activity of FtsZ stabilising the formation of FtsZ fibres. The ZapA protein is widely found in bacterial species and is a relatively small protein which can be expressed in high yield making it ideal for specialised NMR labelling and biochemical studies. In this project, I intent to address important biological questions including: - What is the interaction site between FtsZ and ZapA? - How does the binding of ZapA to FtsZ mediate its effect on FtsZ polymerisation? - How do other binding partners including ZipA, ZapB and FtsA effect this interaction?

16 MicroRNAs in brown and white adipocytes Federica Dimitri Christian Group WMS, Cell & Developmental Biology, Biomed Sci

Two types of adipose tissue exist: white (WAT) and brown (BAT) adipose tissues. WAT stores energy while BAT consumes fatty acids and produces heat by non- shivering thermogenesis through Uncoupling Protein 1 (UCP1). They cooperate in maintaining the energy homeostasis balance and understanding their physiol- ogy is important for the development of treatments against diseases where this equilibrium is compromised, such as obesity and associated metabolic disorders. MicroRNAs (miRNAs) are small non coding RNAs regulating gene expression at the post-transcriptional level and an increasing body of evidence suggests their involvement in adipogenesis and adipose metabolism. Through microRNA array we identiﬁed miRNAs differentially expressed between human mature white and brown adipocytes. MiRNAs can also be secreted into the extracellular environment and be taken up by distal cells, mediating cell-to-cell communication. However, very little is known about adipose tissue-derived circulating miRNAs. Through miRNA PCR array analysis we identiﬁed several miRNAs that are differentially secreted among mouse undifferentiated and differentiated brown and white adipocytes. Bioinformatics target prediction revealed that these miRNAs are potentially involved in important processes regulating the functioning of adipose tissue and its cross-talk with distal cells. Among these, miR-196a showed a con- servative pattern of secretion and intracellular expression in different adipocyte models.

17 How does the Arabidopsis auxinome respond to auxins applied as herbicides? Martin Kubes Napier Group School of Life Sciences

Auxin as a plant hormone plays an irreplaceable role in control and regulation of plant growth and development. This phenomenon led to the idea of synthetic auxins, which remain amongst the most widely used class of herbicides after more than 70 years of application. Generally, they are used to control broad-leaved (dicot) weeds in cereal (monocot) crops in terms of competition for space, light, water and nutrients. A broadly accepted but still hypothetical mechanism of their action says that the synthetic auxin accumulates in plant organs and overloads the endogenous auxin response system, leading to overproduction of e.g. ethylene biosynthetic enzymes or activation of the abscisic acid stress response pathway, and a cascade of downstream responses finally results in plant death. But, if this mechanism is true, why are monocots tolerant to auxin herbicides? One possible explanation is that monocots and dicots manage their responses to high auxin doses differently. Here in the first step we want to use state-of-the-art analytical chemistry techniques to evaluate the responses of dicots to a natural and a synthetic auxin, monitoring the auxinome over time after treatment and evaluating probable mechanisms of action and resistance complemented by genetics, biochemistry, molecular biology and informatics. Preliminary data in Arabidop- sis thaliana, our dicot plant model, clearly support the idea that synthetic auxins not only switch on native response pathways, but that the plants fight back strenuously. Also, work is illustrating which auxin homeostatic pathways domi- nate under individual auxin herbicide pressure.This work was supported by the EU MSCA-IF project CrysPINs (792329).

18 Aging Characteristics of Lithium-ion Batteries Selcuk Atalay Energy Storage Group WMG

As the lithium-ion batteries become increasing prevalent in daily-life and industry, such as electric cars, aircrafts and portable devices, monitoring and predicting battery aging becomes critical. Typically, aging occurs due to multiple complex phenomena and reactions that are occurring simultaneously at different places in battery, and the degradation rate varies between different certain stages during a load cycle, temperature. Battery aging is found to be linear in the early stage of cycling, however, highly non-linear at the end with rapid capacity drop. In this work, we investigate this effect, by systematically account the mass and charge transport with interfacial reactions and parasitic solid-electrolyte-interface (SEI) and lithium plating forming reactions at the anode electrode surface as a function of porosity, C-rate and depth-of-discharge. Our ﬁndings are expected to improve the understanding of battery aging and we expect this fundamental model formulation to be promising for future predictive battery studies and design optimization.

19 A machine learning-developed classiﬁcation system of GP practices, and its relationship to antibiotic prescribing. Saran Shantikumar Communicable Disease Control Evidence and Epidemiology (CDCEE) Group WMS, Population, evidence & technology, Health Sci

Background and Aim: The ability to identify practices with inappropriately high antibiotic prescribing is crucial in determining where to target interventions. Us- ing the same threshold to define high prescribing for all practices disregards other important features of the practice population, such as comorbidity prevalence. Here, we employ a machine learning (ML) approach to classify GP practices by multiple demographic variables and comorbidities; and explore whether the re- sulting classification is associated with antibiotic prescription rates. Methods An unlabelled dataset of all primary care practices in England was curated to include: QOF-reported disease prevalence, smoking and obesity prevalence, index of multiple deprivation score, list size, age and sex-structure (2017/2018). After determining the optimum number of clusters to partition the data (gap statistic), unsupervised ML by K-medioid clustering determined a classification of practice types. The distributions of antibiotic prescribing in each cluster were compared using the Kruskal-Wallis test, with post-hoc Bonferroni-corrected Mann-Witney tests. Results: Five clusters of GP practice were defined, broadly distinguish- able by their overall prevalence of chronic disease, level of deprivation and age distribution. The clusters demonstrated geographical patterning throughout Eng- land. Each cluster had a significantly different distribution of antibiotic prescribing rates (p < 0.001 for all), suggesting that the use of a standard threshold for “high” prescribing may result in misclassification. Discussion: Machine learning success- fully identified clusters of practices with similar characteristics. This clustering approach is useful in comparing practice prescribing rates to other practices with similar characteristics; indeed, it more accurately identified outliers of antibiotic prescribing.

20 Analysing computing performance through modelling and proxy applications James Davis High Performance Scientiﬁc Computing Group Department of Computer Science

The rise of scientific computing and simulation has seen a great increase in the demand for computing resources, using models to simulate real-world systems in areas such as weather-prediction, fluid dynamics and more. This is especially true in the High Performance Computing (HPC) domain, where simulations share the workload of individual large problems across distributed or parallel computing devices. Examples include the use of cluster architectures, where CPUs are spread across many hosts and share their results via a high-speed network, or GPUs, which are effective at repeating the same computational operation across large data-sets in parallel. This enables the investigation of more complex problems of interest in a reasonable timeframe. The difficulty of these approaches however is the complexity involved in ensuring their efficient operation. They must not only be optimal algorithmically, they frequently must also consider the architectures upon which they will be run. By introducing many potential performance bottlenecks such as CPU, memory, network etc., it becomes easy to under- utilise a machine, leading to lower scientific throughput due to the longer wait for results. Our research is in the domain of performance prediction and analysis. The goal of performance modelling is to capture the behaviour of an application in a mathematical model or simulation, linking the cost/time to run them to the characteristics of a machine such that it can be suitably predicted. The use of mini and/or proxy-apps help facilitate this, by reducing the complexity of what might be very large and prohibitively complex codebases to the key algorithms and behaviours that influence their performance. By combining the two areas, this enables us to help identify both what machines are best to use currently, and prepare for architectures of the future.

21 The potential of an ancient remedy for the treatment of diabetic foot infections Blessing Anonye Harrison Group School of Life Sciences

The rise in antimicrobial resistance has prompted the use of alternatives such as plant derived compounds. Previous research from the lab showed that a mixture of plants such as onions, garlic combined with wine and bovine bile had efficacy against Staphylococcus aureus grown in an in vitro model of soft tissue infections (Harrison et al., 2015). Furthermore, this remedy also had efficacy against methicillin-resistant S. aureus (MRSA) in an in vivo mice model (Harrison et al., 2015). The current research focussed on determining the activity against different bacterial species (aerobic and anaerobic) affecting individuals with diabetic foot infections. To do this, the ancient remedy was prepared, mixed together and kept in the fridge for 9 days as specified in the recipe before being tested. In addition, as the research is geared towards developing a topical treatment for diabetic foot infections, several dilutions of the remedy were tested using cell viability and toxicity assays against mammalian cell lines. The results demonstrated that in planktonic culture, the remedy had activity against a range of Gram-negative and positive organisms. However, using the in vitro collagen model for biofilm, it mostly exhibited activity against several Gram-positive organisms including the MRSA strain. The cell viability results showed that lower dilutions did seem to affect the viability, but higher dilutions of the remedy had no effects on skin and immune cells. This was performed through direct contact with cells, but plans are underway to optimise these experiments. This research has the potential to transform the management of diabetic foot infections. Reference Harrison, F., A. E. L. Roberts, R. Gabrilska, K. P. Rumbaugh, C. Lee and S. P. Diggle (2015). ”A 1,000-Year-Old Antimicrobial Remedy with Antistaphylococcal Activity.” mBio 6(4): e01129-01115.

22 In situ Operando Characterisation of Lithium Ion Battery Degradation Anisha Patel Electrochemical Materials Group WMG

As rechargeable batteries become an integral part of our lives and our dependence on them continually increases, extensive research efforts have been invested in improving existing battery components. However, batteries are complicated systems in which a variety of chemical, electrochemical and physical processes take place within dynamic microstructures. Greater fundamental understanding within these processes is critical for optimising the battery design and improving battery life and efﬁciency. One approach is to couple a powerful microscopic characterisation technique with electrochemical operation. This enables critical studies around capacity fading issues, such as dendrite formation and physical degradation under realistic conditions. Direct, real-time monitoring allows for correlation of physical and chemical processes to the electrochemical responses of a battery, producing powerful information, leading to a better understanding of lithium ion (Li-ion) battery degradation processes. Scanning electron microscopy (SEM) is a powerful microscopic technique that is based on data obtained from an electron beam to construct an image of a surface (down to nm resolution), allowing nanoscale monitoring of microscopic morphology changes. Coupling this with electrochemistry in operando conditions enables correlation of structure, morphology, composition and performance to be acquired. In order to monitor the interface of energy storage materials with the electrolyte in a coin cell battery in real time, a bespoke in situ coin cell designed has been designed in order to visualise the electrode/electrolyte interfaces and capture morphological changes. The cell allows monitoring of degradative processes such as dendrite growth, where the morphology variations and growth extent can be characterised and correlated with electrochemical data. These studies aim to enhance our understanding of the processes taking place that lead to electrode degradation and failure.

23 24 Posters

25 P01: High resolution imaging of conjugated polymers using combined electrospray deposition and scanning tunnelling microscopy Luis Alves Perdigao Costantini Group Department of Chemistry

Electrospray ion beam deposition (ESI-BD) has become a versatile technique for vacuum depositing large thermally labile molecules on surfaces under controlled conditions and has been applied to molecular magnets, biomolecules, and por- phyrin nanorings. In this talk, by combining vacuum electrospray deposition and high-resolution scanning tunnelling microscopy (STM) we demonstrate imaging of conjugated polymers with unprecedented detail, thereby unravelling structural and self-assembly characteristics that were previously impossible to achieve. Sub- monomer resolution STM images of polymers clearly reveal the conjugated back- bones and the solubilising alkyl side-chains in individual polymer strands. Based on this, it becomes possible to determine the molecular number distribution of the polymer by simply counting the repeat units. More importantly, we demonstrate that we can identify, precisely determine the nature, locate the position, and ascertain the number of polymerisation defects in the backbone of diketopyrrolopyrrole (DPP)-based polymers. This unique insight into the structure of conjugated polymers is not attainable by any other existing analytical technique and represents a fundamental contribution to the long-discussed issue of defects as a possible source of trap sites. Furthermore, the analysis of our high-resolution images, also reveals that the frequently assumed all-trans-conformation of the monomers in the polymer backbone is actually not observed, while demonstrating that the main driver for backbone conformation and hence polymer microstructure, is the maxi- mization of alkyl side-chain interdigitation. In contrast, imaging studies of IDT-BT polymers reveal that their synthesis produces mainly defect-free sequences, and their measured length is below the values obtained by GPC analysis.

26 P02: First-Principles Molecular Dynamics Study of Rechargeable Lithium Based Batteries Shalini Awasthi WMG

Due to increase in global energy demand, developing energy storage systems with higher energy densities are an urgent need. Rechargeable Li-ion batteries (LIBs) emerge as an important power sources for energy storage device for portable electronics and are being continuously investigated and developed for use in automo- bile electric vehicles (EV). However, during the cycling of LIBs, due to the decomposition of electrolyte, a passivation layer called solid electrolyte interphase (SEI) is formed on electrode surface. Although, SEI formation is vital in LIBs for long term performance, it leads to irreversible loss of lithium and thus, storage capacity of battery. While several studies have been performed in the past few decades on SEI, it is difﬁcult to control its formation and growth, as the chemical composition, and stability of SEI depend on a list of factors. These factors include the electrode material, electrolyte composition, electrochemical conditions, and cell temperature. The perfect SEI would be a fast forming, ﬂexible, stable, and contains insoluble species with low electronic and high ionic conductivity. Thus, the formation of SEI and the electrochemical stability of SEI components should be a primary topic of investigation in future development of LIBs. To understand the stability and chemistry of SEI layer, density function theory based molecular dynamics simulations were performed. We used enhanced sampling techniques to understand various degradation pathways of electrolyte decomposition in LIBs in various reaction conditions as well as in presence of additives.

27 P03: Surveying atoms inside nanometer scale wires Aruni Fonseka Microscopy Group Department of Physics

The constant drive for smaller, smarter and more efficient electronics has led to miniaturisation of electronic devices such as transistors, enabling packaging of millions of such devices into a single chip. However, miniaturisation is now reaching the limits defined possible by laws of physics. This has led to the evolution of alternative solutions, one of which is making use of other spatial dimensions and stacking devices on top of each other, and hence reducing the foot-print per device. One fundamental structure that makes this possible are the semiconductor wires that have nanometer scale diameters and micrometer scale lengths. These are commonly known as nanowires. Here, the devices can be assembled along its length and on its sides while maintaining a very small footprint on the chip. Nanowires have already been demonstrated in many electronic applications such as transistors, sensors and LEDs. Like in any other field, understanding of their morphology and structure is paramount for the advancement of the nanowire research area. However, at nanometer scale, one is effectively probing distances that are in fact only few atoms in width. As this length scale is far smaller than the diffraction limit of light, scientists have resorted to microscopes that use a medium that has a much smaller wavelength than light, and in this case, it is the wave nature of electrons in a transmission electron microscope (TEM). This presentation will discuss some of the atomic scale TEM analysis done at Warwick microscopy facility on semiconductor nanowires. One of the main findings is the unexpected self-formations that take place within the nanometer-scale diameter of seemingly plain nanowires. The high-resolution analysis has been able to identify down to the level of atom columns of individual elements, forming a clear understanding of the structure of nanowires and hence providing answers to otherwise inexplicable observations made by other characterisation techniques.

28 P04: Uterine natural killer cells are regulated by vascular adhesion protein-1 in the endometrium Seley Gharanei Zammit Group WMS, Translational & Experimental Med, Biomed Sci

Leukocyte extravasation into tissues is mediated by adhesion molecules expressed on endothelial cells and pericytes. Here we show that pericytes around the spiral arterioles in midluteal human endometrium constitutively express vascular adhesion protein-1 (VAP-1), a transmembrane protein involved in immune cell traf- ficking and inflammation. Knockdown of VAP-1 impaired the invasive, contrac- tile, migratory and adhesive capacities of primary endometrial pericytes. Several lines of evidence indicated that VAP-1 is a pivotal gatekeeper for uterine natural killer (uNK) cells required for optimal hemochorial placentation in both human and mouse. First, the abundance of CD56+ uNK cells in midluteal human endometrium correlated positively with the expression of AOC3, encoding VAP-1. Second, Aoc3 deletion in mice resulted in complete absence of Dolichos biflorus agglutinin-positive uNK cells during pregnancy. Finally, we show that loss of Aoc3 reduces litter size in mice, whereas low AOC3 expression in human endometrium is associated with recurrent pregnancy loss.

29 P05: Mathematical modelling of dilute granular gases Vinay Kumar Gupta Micro and Nano Flows Group Warwick Mathematics Institute

A collection of discrete macroscopic particles which dissipate energy during colli- sions among themselves is termed as a granular material. Granular materials are ubiquitous in nature and industry alike. They occur in all shapes and sizes ranging from few microns to several hundred kilometres. Sand dunes, debris flow, aster- oid belt, dust storm, gravels, cement, food grains, sugar, capsules, pills are some typical examples of granular materials. Depending on the energy supplied, they can exist in the solid, liquid or gaseous state. Due to their dissipative nature, granular materials exhibit several interestingand often counter-intuitivephenomena. At the same time, this very feature of granular materials poses many difficulties while modelling processes in granular materials due to the non- conservation of energy. The present work considers the gaseous state of granular materials, for which, analogous to molecular gases, mathematical tools can be developed within the framework of kinetic theory. In this talk, I shall present my recent work on moment-based mathematical models for modelling a dilute granular gas of d-dimensional smooth, identical, inelastic spheres interacting with Maxwell interaction potentialreferred to as inelastic Maxwell molecules (IMM). Here d = 2 means disk flows while d = 3 means sphere flows. This work is a somewhat generalization of my previous work [1] to arbitrary dimensions. To assess the capabilities of the derived models for IMM, the homogeneous cooling state of a freely cooling granular gas of IMM and its stability to small perturbation is stud- ied. Moreover, the Navier-Stokes level transport coefficients are obtained from the moment equations and it has been found that they agree exactly with those obtained at the first-order Chapman-Enskog expansion of the Boltzmann equation [2]. References: [1] V. K. Gupta, P. Shukla & M. Torrilhon, J. Fluid Mech., 836, 451501 (2018). [2] A. Santos, Physica A, 321, 442466 (2003).

30 P06: Can we end the vicious cycle of antibiotic resistance in CF? Marwa Hassan Harrison Group School of Life Sciences

Bacterial biofilms are a complex structure of bacterial cells, extracellular matrix and DNA that act together to protect the bacterial biofilm from antibiotics and immune cells. Thus, bacterial biofilms are known to have high antibiotic tolerance which directly affects clearance of bacterial infections in cystic fibrosis patients. Current antibiotic susceptibility testing methods are either based on planktonic cells or do not reflect the complexity of biofilms in vivo. Consequently, inaccurate diagnostics affect the treatment choice, preventing bacterial clearance and developing antibiotic resistance. This leads to prolonged treatment especially in cystic fibrosis. In this study, we are using an ex-vivo biofilm lung model to study antibiotic tolerance in cystic fibrosis to develop better diagnostics. Sections of pig bron- chiole were prepared as previously described, infected with lab strains and clinical isolates of Staphylococcus aureus or Pseudomonas aeruginosa and incubated in artificial sputum media to form biofilms. Then, lung-associated biofilms were challenged with antibiotics and their bacterial load was quantified. All isolates were also tested for antibiotic susceptibility using standard planktonic and biofilm methods. The results showed increased antibiotic tolerance/resistance of both S. aureus and P. aeruginosa against all tested antibiotics. Thus, all sensitive bacterial isolates, according to the standard antibiotic susceptibility testing, demonstrated a resistant phenotype in the ex-vivo biofilm lung model. We demonstrate a realistic model for antibiotic susceptibility testing clinically and in anti-biofilm drug development to help understanding antibiotic resistance and tolerance in biofilms.

31 P07: Hidden Markov models for monitoring circadian rhythmicity in telemetric activity data Qi Huang Cancer Chronotherapy Team WMS, Translational & Experimental Med, Biomed Sci

Wearable computing devices allow collection of densely sampled real-time information on movement enabling researchers and medical experts to obtain objective and non-obtrusive records of actual activity of a subject in the real world over many days. Our interest here is motivated by the use of activity data for evaluating and monitoring the circadian rhythmicity of subjects for research in shift-workers and chronotherapeutic healthcare. In order to translate the information from such high-volume data arising we propose the use of a Markov modelling approach which (i) naturally captures the notable square wave form observed in activity data along with heterogeneous ultradian variances over the circadian cycle of human activity, (ii) thresholds activity into different states in a probabilistic way while respecting time dependence, (iii) summaries the average day oscillations by introducing periodic harmonics and (IV) gives rise to circadian rhythm parameter estimates, based on probabilities of transitions between rest and activity, that are interpretable and of interest to circadian research.

32 P08: Macrocyclisation of Small Peptides Enabled by Oxetane Incorporation Stefan Roesner Shipman Group Department of Chemistry

Compared to their linear counterparts, cyclic peptides benefit from enhanced cell permeability, increased target affinity, and higher resistance to proteolytic degradation. Moreover, they are capable of acting as inhibitors against some of the most challenging targets. One major obstacle to the discovery and development of new cyclic peptide drugs relates to the difficulties associated with their synthesis. The head-to-tail cyclisation of short peptides containing seven or less amino acids is especially challenging. Common problems encountered during cyclization are C- terminal epimerisation, cyclooligomerisation and the appearance of side products arising from polymerisation. Consequently, there is a pressing need to discover new macrocyclisation strategies that can provide easy access to a variety of small cyclic peptide scaffolds. A new type of turn-inducing element for peptide macrocyclisation in which one of the backbone C=O bonds is replaced with an oxetane ring is reported. The cyclisation precursors are conveniently made in solution or by solid-phase peptide synthesis. The subsequent macrocyclisations are demonstrated to be general, enabling a variety of difficult ring closures including those producing head-to-tail and side-chain to side-chain linkages. The method tolerates variation in the size of the macrocyclic ring and the location of the turn-inducing 3-amino-oxetane element with respect to the amide bond being formed. Addi- tional data reveal that oxetane modified cyclic peptides are excellent bioisosteres of conventional cyclic peptides.

33 P09: Phospho-regulation of tropomyosin is crucial for actin cable turnover and division site placement Saravanan Palani Balasubramanian Group WMS, Cell & Developmental Biology, Biomed Sci

Tropomyosin is a coiled-coil actin binding protein key to the stability of actin filaments. Whereas, in muscle cells tropomyosin is subject to calcium regulation, how it is regulated in non-muscle cells is unknown. Here, we provide evidence that the fission yeast tropomyosin, Cdc8, is regulated by phosphorylation of a ser- ine residue. Failure of phosphorylation leads to an increased number and stability of actin cables and causes misplacement of the division site in certain genetic back- grounds. Phosphorylation of tropomyosin weakens its interaction with actin filaments without affecting its structure or stability. Furthermore, we show through in vitro reconstitution that phosphorylation-mediated release of tropomyosin from actin filaments facilitates access of the actin severing protein cofilin and subsequent filament disassembly. These studies establish that phosphorylation may be a key mode of regulation of non-muscle tropomyosins, which controls in fission yeast actin filament stability and division site placement.

34 P10: Accumulation of ferric iron in ageing Drosophila brain is independent of macroautophagy Anne-Claire Jacomin Nezis Group - Autophagy in Health and Diseases School of Life Sciences

Biometals such as iron, copper, potassium and zinc are essential regulatory elements of several biological processes. The homeostasis of biometals is often affected in age-related pathologies. Notably, impaired iron metabolism has been linked to several neurodegenerative disorders. Autophagy, an intracellular degradative process dependent on the lysosomes, is involved in the regulation of ferritin and iron levels. Impaired autophagy has been associated with normal, patho- logical ageing and neurodegeneration. Non-mammalian model organisms such as Drosophila have proven to be appropriate for the investigation of age-related pathologies. Here, we show that ferritin is expressed in adult Drosophila brain and that iron-bearing ferritin accumulates in the brain of old wild-type flies but not in autophagy-deficient fly brains. To further investigate the level and state of iron in the brain, we made use of synchrotron Xray fluorescence microscopy (µXRF). Our study revealed that the accumulation of iron in ageing Drosophila brain corresponds to ferric iron oxide typical of that found in ferritin.

35 P11: Mechanism of biological effect of the new photosensitive organometallic iridium(III) complexes Jitka Pracharova Sadler Group Department of Chemistry

Despite the clinical success of platinum-based chemotherapeutics, significant research efforts have focused on developing compounds based on other metals1. In recent years, there has been particular interest in the development of ruthenium, osmium, and iridium based drugs and several have been shown to exhibit anticancer activity in vitro and in various animal models2-4. In this work, the cytotoxic potential and mechanism of biological effect of a new luminescent organometallic iridium(III) complexes of the type [Ir(CN)2(Nˆ N)][PF6]ˆ (CNˆ = 2-(paratrifluoro- methyl-phenyl)-benzimidazol-N,-C; NNˆ = 1,10-phenanthroline, 1; dppz-izdo = dipyrido[3,2-a:2,3-c]phenazine-10,11-imidazolone, 2) were investigated. Both 1 and 2 showed promising cytotoxicity in all cell lines tested including cisplatin- resistant A2780R cells. The Ir complexes 1 and 2 were also less cytotoxic in non- cancerous breast cells, MCF-10A and exhibited higher selectivity for breast cancer cells MCF-7 than conventional platinum anticancer drug, cisplatin. The investigated Ir complexes induce apoptosis and possible mechanism of their action involves inhibition of protein translation. Furthermore, it was found that biological effect of compounds 1 and 2 can be markedly enhanced by irradiation with visible light. In conclusion, a representative 1 and 2 iridium complexes seem to have interesting biological properties and are candidates for use in photoactivated cancer chemotherapy. 1 S. Dasari, PB. Tchounwou, European Journal of Pharmacology, 2014, 740, 364. 2 S. H. Rijt, H. Kostrhunova, V. Brabec, P. J. Sadler, Bioconju- gate Chemistry, 2011, 22, 218. 3 A. Bergamo, G. Sava, Chemical Society Reviews, 2015, 44, 8818. 4 J. Yellol, S. A. Pérez, G. Yellol, J. Zajac, A. Donaire, G. Vigueras, V. Novohradsky, C. Janiak, V. Brabec, J. Ruiz, Chemical Communications, 2016, 52, 14165.

36 P12: Tracking kinetochores in lattice light sheet microscopy data Jonathan Harrison Zeeman Institute (Burroughs and McAinish groups) Warwick Mathematics Institute

Lattice light sheet microscopy is revolutionizing cell biology since it enables fast, high-resolution extended imaging in three dimensions combined with a drastic reduction in photo-toxicity/bleaching [1]. However analysis of such data sets still remains a major challenge. Automated tracking of particles in imaging data can reveal quantitative insights into biological mechanisms, such as those inﬂuencing the mechanics of kinetochores, the protein complex facilitating and controlling microtubule attachment of the chromosomes within the mitotic spindle. Here we extend existing open-source kinetochore tracking software (KiT [2]) to track (and pair) kinetochores throughout pro-metaphase to anaphase in lattice light sheet microscopy data. This involves implementing frame-to-frame alignment and dealing with the size of the imaging ﬁles. This software provides quantitative insights into how kinetochores robustly ensure capture, bi-orientate and congress chromosomes during mitosis. [1] Chen, B. C. et al. (2014) Science, 346(6208), 1257998. [2] Armond, J. W., Vladimirou, E., McAinsh, A. D., & Burroughs, N. J. (2016) Bioinformatics, 32(12), 1917-1919.

37 P13: Aberrant Speciﬁcation Of Endometrial Stromal Cells Into Distinct Decidual Subpopulations In Recurrent Pregnancy Loss. Emma Lucas Reproductive Health Group WMS, Cell & Developmental Biology, Biomed Sci

Pregnancy depends on successful transformation of the cycling endometrium into a semi-permanent tissue, the decidua, maintained throughout gestation. While this process is orchestrated by ovarian and embryonic cues, it is effected by the dif- ferentiation and polarization of endometrial stromal cells (EnSC) into two subpopulations: mature stress-resistant and acutely senescent decidual cells. Mature decidual cells then cooperate with uterine NK cells to eliminate their stressed/senescent counterparts, thereby curtailing tissue inflammation and enabling transfiguration into a gestational tissue. Aberrant decidualization causes recurrent pregnancy loss (RPL) but the underlying mechanisms remain poorly understood, in part due to lack of markers for decidual subpopulations. To overcome this hurdle, we mapped the temporal transcriptomic changes in primary EnSCs along a decidual time- course and in seven endometrial biopsies, spanning the implantation window, using scRNASeq. The expression of mature and senescent decidual marker genes was measured by RT-qPCR in biopsies from RPL and control subjects. Decidualization in vitro starts with a precipitous transcriptional response, followed by synchronous transition of EnSCs through intermediate states before emerging as divergent subpopulations, representing mature and stressed decidual cells. A total of 326 DEG were identified between the decidual subpopulations. In vivo, transition from receptive to post-receptive endometrium was also marked by progression of En- SCs along diverging transcriptional trajectories, involving genes with conserved branching dynamics in vivo and in vitro. RPL was associated with significant reduction in mature decidual marker genes and increased expression of senescent genes. These data suggest that aberrant specification of EnSCs into functionally distinct decidual subpopulations during the implantation window predisposes for subsequent pregnancy loss.

38 P14: Control of sp2 content in boron doped diamond for quantitative detection of hypochlorite at high concentrations Anthony Lucio Macpherson Group Department of Chemistry

An electrochemical based sensor that contains selectively patterned regions of non-diamond carbon (i.e. sp2) in a boron-doped diamond (BDD) matrix is presented for the quantitative detection of hypochlorite (OCl-) at high concentrations in highly alkaline, chemically oxidising solutions. By using a laser micromachining process it is possible to mechanically write robust regions of sp2 carbon into the principally sp3 BDD electrode, and we ﬁnd that these laser machined regions are active towards the hypochlorite ion reduction process. In this work we examine four different laser patterned BDD electrodes and compare their response across a range of hypochlorite ion concentrations (i.e. 0.02 M to 1.50 M OCl-). Experimen- tal electrochemical data are complimented with ﬁnite element modelling (FEM) data to better understand the electrochemical system as contributions from migra- tion to mass transport are expected to occur at these high concentrations. Using a closely spaced sp2 micro-spot array, where diffusional overlap between different pits occurs, a linear response towards OCl- from square wave voltammetry data is obtained within the concentration range from 0.02 M to 1.50 M OCl- (1500 ppm to 112000 ppm; sensitivity = -0.127 0.004 mA M-1; R2 = 0.992). FEM simulation data suggest that only a small percentage of the total laser machined sp2 area is actually active towards the hypochlorite reduction process. Nonetheless, these types of hybrid sp2-sp3 diamond-based electrochemical sensors show great promise for use in bleaching applications ranging from paper pulp processing to detoxifying metal plating baths.

39 P15: Chasing gravitational wave sources in the electromagnetic domain Joseph Lyman Astrophysics Group Department of Physics

The culmination of Einstein’s century-old prediction of gravitational waves, as part of his General Relativity framework for how the Universe works, occurred in 2015 with the first direct detection of gravitational waves. Gravitational waves arise as ripples in space-time due to accelerating masses, but were thought by Einstein to be to miniscule to observe. Following many decades of engineering, the LIGO gravitational wave detectors managed to clearly detect this first signal, and determined it to have arisen from a merging binary black hole system over a billion light years away. Shortly after the birth of this entirely new way to observe the Universe, in August 2017, a new gravitational wave signal was found, arising from the merger of two neutron stars, known as GW170817. Neutron star mergers had long been predicted to also produce electromagnetic radiation (unlike black hole mergers), which thus prompted a significant fraction of the astronomical com- munity to quickly search the sky for the counterpart to this gravitational wave signal. Following the successful search for, and analysis of, this landmark ’multi- messenger’ event, I will present the huge impact it has had on astrophysics and cosmology, detail the exciting developments now that we have opened our eyes (photons) and ears (gravitational waves) on the Universe, and look at our own Warwick-led observatory - GOTO - that will be chasing future gravitational wave events.

40 P16: Analyses on Representations of Materials for Machine Learning Berk Onat Kermode Group School of Engineering

Machine learning has received increasing attraction in materials science, chemistry, physics and biology as the need for extracting and classifying materials and molecules becomes daunting as a result of the large amount of experimental and theoretical data that is now available. Descriptors for materials and molecules based on physical, chemical and biological properties have been used in machine learning approaches in these fields so far. However, descriptors with efficient structural representations based on atomic configurations are also introduced recently such as atom-centred symmetry functions (ACSF) [1], Smooth Overlap of Atomic Positions (SOAP) [2] and Many-Body Tensor Representation (MBTR) [3]. While these recent descriptors are becoming a common practice in machine learning applications, the sensitivity of their structural representations and how capable they are covering the structural phase space of materials are not fully analysed. In this presentation, we will discuss the classification of these representations and how one can examine their sensitivity to changes in atomic configurations using large datasets. We will present our results on the analyses of the performance of various representations of materials on bulk and molecular datasets extracted from NOMAD European Centre of Excellence archive[4]. [1] J. Behler and M. Par- rinello, PRL 98, 146401 (2007).[2] A. P. Bartok, R. Kondor, and G. Csanyi, PRB 87, 184115 (2013).[3] H. Huo, M. Rupp, arXiv 1704.06439, (2017).[4] The Novel Materials Discovery Laboratory - A European Centre of Excellence, http://nomad- coe.eu

41 P17: Modelling rare-earth/transition-metal magnets at ﬁnite temperature with self-interaction-corrected relativistic density-functional theory and disordered local moments Christopher Patrick Theory Group Department of Physics

Intermetallic rare-earth/transition-metal (RE-TM) compounds can show excep- tionally hard magnetic properties, the most famous examples being neodymium iron boride (Nd-Fe-B) and samarium cobalt (Sm-Co). Such compounds present an interesting challenge to theory: their magnetization and high Curie temperatures derive from the itinerant electrons of the transition metal, generally quite well described by the local spin-density approximation (LSDA) to density-functional theory, but their magnetocrystalline anisotropy originates from the highly local- ized 4f electrons of the rare earth, which are not well accounted for by the LSDA. Here we present a scheme to calculate the ﬁnite temperature magnetic properties of RE-TM compounds which combines relativistic disordered local moment theory with the local self-interaction correction, with the latter used to improve the LSDA description of the 4f electrons. We demonstrate the scheme by calcu- lating the magnetization vs. temperature curves and Curie temperatures (Tc) of the RECo5 class of RE-TM magnets, RE = Y–Lu. We show how the scheme repro- duces experimentally-measured trends across the lanthanide series. We also use an order parameter analysis to investigate the effect of the RE on the Co-Co interactions, which strongly inﬂuence Tc. We tentatively attribute the variation in these interactions with RE to a small contribution to the density from f-character electrons located close to the Fermi level.

42 P18: Symmetry design of Magnetoelectric coupling in Perovskites Fernando Pomiro Senn Group Department of Chemistry/Institute of Advanced Study

Multiferroics are a promising class of functional materials for data storage and sensing applications. In these materials, magnetic and ferroelectric orderings co- exist. However, the coexistence of these two kinds of order in the same material does not necessarily imply that they are strongly coupled with each other, such as is required if the reversal of magnetisation is to affect a reversal of the direction of ferroelectric polarisation. Such coupling is vital if these materials are to find util- ity in next generation hard drives where storage bits will be written electronically but read magnetically. In this work, a group-theoretical approach is used to enu- merate the possible couplings between magnetism and ferroelectric polarization in the parent Pm-3m perovskite structure, or in other words, we seek to generalize a recipe for inducing magnetoelectricity in the perovskite structure. These recipes are based on symmetry arguments alone, and the ingredients are structural and magnetic degrees of freedom, which are classified in terms of transforming as ir- reducible representations of the parent space group. Both the rich chemistry of the perovskite structure, and their 3D nature compared to the Ruddlesden-Popper phases, make for a superior play-group to explore these magnetoelectric coupling, and will ultimately allow for the design of a room temperature multiferroics with a strong magnetic coupling that may be easily processed into thin films and devices.

43 P19: Characterisation of the Circadian Modulation of Blood-Brain-Barrier Permeability in an In Vitro Tri-Culture Mode Swati Kumar Chronotherapy Group WMS, Translational & Experimental Med, Biomed Sci

Brain penetration of most molecules is limited by the Blood-Brain-Barrier (BBB), which has evolved to protect the brain from harmful substances and pathogens, yet allows others to enter. This is highly relevant for drugs targeting the central nervous system (CNS), but also drugs with considerable CNS toxicities, such as many chemotherapeutics. The BBB is established through tight junctions formed mainly by claudins and occludins between endothelial cells surrounding the brain vasculature, pericytes and other glial cells. There is some but limited evidence suggesting the circadian clock modulates BBB function in vivo. An in vitro BBB model would enable cost effective and 3Rs conform screening of drugs for iden- tification of maximum and minimum BBB permeability, which could be used to improve drug levels in the brain or reduce toxic effects on the brain, respectively. Here, we describe circadian characteristics of an in vitro tri-culture BBB model containing primary murine astrocytes, pericytes and endothelial cells. First, bio- luminescence oscillations were confirmed using monocultures of the different cell types from circadian clock (PER2::LUC) mice after forskolin (fsk) synchronisation. Then, the barrier tightness of the in vitro triculture BBB was assessed measuring daily trans endothelial electrical resistance (TEER) for two weeks. TEER steadily increased to ca. 260 Ω/cm2 on day 12, and days 8-12 were identified as optimal time window for drug permeability testing. Circadian changes in TEER were assessed after fsk shock on day 8 (0h). Peak TEER levels at 36h were about 25% higher than TEER at 24h and 48h suggesting circadian changes in BBB tightness. This was in line with immunocytochemistry results suggesting about 2-fold higher claudin-5 and occludin expression levels in endothelial cells at 36h compared to 24h. Finally, apparent permeability was measured every 4h for 24h using FITC- dextrans. In anti-phase to the tightness, ca. 3-fold more 4kDa and 10kDa dextran passed through the model at 24h compared to 36h after fsk shock. Importantly, 70kDa dextran did not pass at any time suggesting a well-developed barrier function. Taken together, our data suggest that we have established a tri-culture in vitro BBB model, which displays significant circadian rhythms in tightness and an anti-phasic rhythm in drug permeability. In the future, our model may be used to optimise timing of clinically approved and new drugs with CNS targets or toxicities.

44 P20: A Boron Doped Diamond Multi-Tool Electrode For Simultaneous pH and Dissolved Oxygen Analysis. Tania Read Warwick Electrochemistry and Interfaces Group Department of Chemistry

We describe an electrochemical sensor designed for the simultaneous detection of dissolved oxygen (DO) and pH in buffered aqueous solutions. The sensor com- prises of patterned regions of non-diamond-carbon (sp2) in a boron doped diamond (sp3) matrix, prepared using a laser micropatterning process which makes possible to write mechanically robust regions of sp2 carbon into a BDD electrode. These regions have been demonstrated to both promote the electrocatalytic reduction of oxygen, and facilitate the proton coupled electron transfer of quinone groups integrated into the surface of the sp2 carbon. Consequently, in one voltam- metric sweep (time of measurement 4 s) it is possible to determine both the DO concentration and solution pH. By tailoring the sp2 pattern the response can be op- timised towards both analytes, whilst the carbon based nature of these electrodes offers promise of biocompatibility. Using a closely-spaced sp2 microspot array we have demonstrated a linear response towards DO, across the range 0.0 to 8.0 mg L-1 (0.0 to 0.25 mM; sensitivity = -8.77 × 10−8 A L mg−1,R2 = 0.9991) and pH range 4 10 (sensitivity = 59.7 mV pH−1,R2 = 0.9983). The ability to measure both DO concentration and pH in the more complex buffered environment of blood using these sensors has also been established, offering potential towards fast POC analysis. Finally, it has been found that the peak position for ORR is independent of pH and thus via measurement of the difference in ORR and pH peak position, internal referencing is possible a potential solution to one of the issues suffered by many electrochemical biosensors, reference drift. Although at early stages, such electrodes show great promise for use in applications ranging from bio-medical sensing to water analysis.

45 P21: Sharp bound on the number of maximal sum-free subsets of integers Maryam Sharifzadeh Combinatorics Group Warwick Mathematics Institute

One of the most basic questions in mathematics is “How many?”. In particular, in enumerative combinatorics, one of the classical problems is that of counting the number of elements in a ﬁnite set. The study of these problems have led to the discovery of some exciting and far-reaching methods. Recently, in a sequence of two papers, we completely settled a 30 year-old conjecture of Cameron and Erdos (one of the most proliﬁc mathematicians of the 20th century) in number theory on enumerating maximal sum-free sets of integers. In particular, we gave an asymptotic count of the number of maximal sum-free sets in the set of integers {1,...,n}. This is joint work with Jozsef Balogh, Hong Liu, and Andrew Treglown.

46 P22: Regaining the ashes: ﬁnding chemical ‘clues’ to mitigate the impact of ash dieback John Sidda Grant Group School of Life Sciences

Ash trees, Fraxineus excelsior, are an iconic native British tree comprising 5.5% of our woodland and provide essential habitat for more than 40 additional native UK species. Ash dieback (ADB) caused by the fungus Hymenoscyphus fraxineus, was first reported in Poland in 1992 and has spread westwards, devastating ash populations across Europe. Since the first confirmed reports in UK ash in 2012, the threat ADB poses to the UK landscape has made it the only plant disease to be the subject of two meetings of the UKs emergency COBRA committee.Monitoring the native ash population revealed a small proportion of trees which appear to be tolerant to ADB. These trees show no symptoms in woodlands that are otherwise heavily affected by the disease. By studying the different chemical components (metabolites) of ash trees that are tolerant or susceptible to ADB, we aim to identify specific metabolites which are overproduced by tolerant or susceptible trees. The first class of metabolites to be identified, iridoid glycosides, were shown to be more prevalent in susceptible trees. These metabolites are well-known insect feeding deterrents, therefore by selecting trees which are more tolerant to ADB based on their low iridoid glycoside content, breeders may inadvertently se- lect trees which will be more susceptible to insect herbivores such as Emerald ash borer, which has decimated the ash population across North America in recent years.Through a combination of liquid chromatography-mass spectrometry (LC-MS) and computational analysis we can separate and identify metabolites of different UK ash populations. This approach has identified a complex diversity of discriminatory metabolites; 600 metabolites are produced more frequently in susceptible populations. These results suggest a much greater number of metabolites that may be contributing to ABD susceptibility other than the iridoid glycosides identified in the previous study of Danish trees.

47 P23: Dose enhancement studies in synchrotron radiation therapy Jenny Spiga Medical Physics Group Department of Physics

The local dose deposition obtained in X-ray radiotherapy can be increased by the presence of metal-based compounds in the irradiated tissues; this finding is strongly enhanced if the radiation energy is chosen in the kiloelectronvolt energy range, due to the proximity to the absorption edge. Monte Carlo (MC) calculations facilitate the estimation of this effect by predicting the dose distribution in a specific phantom. In this study, we present a MC application developed with the toolkit GEANT4 to investigate the dosimetric distribution of a uniform monochromatic X-ray beam, and benchmark it against experimental measurements. Two val- idation studies were performed, using a commercial PTW RW3 water-equivalent slab phantom for radiotherapy, and a custom-made PMMA phantom conceived to assess the influence of high atomic number compounds on the dose profile, such as iodine and gadolinium at different concentrations. Good agreement was found for the monochromatic energies considered, which were in the range of 30-140 keV. A dose enhancement was observed in the calculations, corresponding to the regions containing the contrast agents. Dose enhancement factors (DEFs) were calculated, and the highest values were found for energies higher than the corresponding K-edges of iodine and gadolinium. The in-silico results are in line with the empirical findings, which suggest that GEANT4 can be satisfactorily used as a tool for the calculation of the depth dose profiles at low energies in the presence of contrast agents.

48 P24: The Warwick Centre for Ultrafast Spectroscopy: Dynamics, the Full Spectrum Michael Staniforth Stavros/Lloyd-Hughes Group Department of Chemistry

We introduce Warwick’s latest State-of-the-art user facility. Ultrafast (femtosecond - 10−15 s) techniques ranging from the ultraviolet to infrared to terahertz radiation give the Warwick Centre for Ultrafast Spectroscopy a uniquely broad range of wavelengths and experiments available to users from Warwick and be- yond. These ultrafast techniques are complimented by a range of steady-state spectrometers: A Cary60 UV/Vis absorption spectrometer, a Brucker70v FTIR absorption spectrometer, and a Fluorolog ﬂuorescence spectrum that can measure photoluminescence lifetimes on the order of nanoseconds (10−9 s). The facility has already been used to study a wide range of different samples from diverse dis- ciplines such as: Protein tagging, electrochemical luminescence, perovskite materials, sunscreen photo-protection, dye-surfactant interactions, etc. In the next few years we are hoping to add polarisation control in the THz regime, a femtosecond stimulated raman spectroscopy set-up and expand our abilities in time resolved luminescence. If you are interested or think we can help with your own research project, or even if you are unsure and need some clariﬁcation, please contact Dr Mick Staniforth at [email protected] or go to go.warwick.ac.uk/WCUS

49 P25: Bacteriophage: Potentials for enhancing antibacterial therapy Eleanor Townsend Jameson Lab School of Life Sciences

Bacteriophage are viruses that infect bacteria, also known as phages. They are ubiquitous in nature with an estimated 1x10 8 phage per mL of sea water and approximately 10 15 phage particles in the human gut. Phages are able to cause either lysogenic (phage integrates into host genome) or lytic (results in host cell death upon release of progeny phage) infections of bacteria. The tail fibres of phage have also been shown to have depolymerase effects against bacterial cap- sule and biofilm matrix. Klebsiella are a major cause of nosocomial infections, and whilst a normal resident of the human gut, their metabolites can be detri- mental to cardiovascular health. Klebsiella are an ESKAPE pathogen, a group of high-risk and increasing antibiotic resistance. They are also proficient biofilm formers, a phenotype which is well known to increase tolerance to antimicrobial drugs. Lytic phages provide an alternative method of selectively targeting these species without any risk of development of resistance mechanisms. Furthermore, non-lytic phages may provide effects that sensitise bacteria to antibiotics, allowing them to be used synergistically. Our aim is to investigate synergistic and additive potential for bacteriophages in adjunct with antibiotic treatments to disrupt Kleb- siella bacterial communities. We have employed standardised methods of chequer- board minimum inhibitory concentration (MIC) testing, with added time course readings, so that phage-bacteria dynamics can be assessed. A number of antibiotics (ampicillin, vancomycin, and meropenem) have been used alongside phage KppS2, against four unique strains/species of Klebsiella. Results have shown that phage and antibiotics work in conjunction to disrupt planktonic bacterial growth, even in strains of Klebsiella that have limited sensitivity to phage KppS2, high- lighting its potential use in therapy. Future work will transfer these findings into biofilm models, where cells are typically recalcitrant to treatment.

50 P26: Transparent metal electrodes for organic solar cells using an efﬁcient lift-off process Silvia Varagnolo Hatton Group Department of Chemistry

This talk will present a new method for the fabrication of optically thin metal ﬁlm electrodes with ordered arrays of tiny apertures, suitable as transparent electrodes in high performance organic solar cells. This new approach to electrode fabrication uses micro-contact printing to deposit tiny amounts of resist that can be easily lifted-off after metal deposition by vacuum evaporation, leaving clean holes in the ﬁlm. The potential of these electrodes as transparent electrodes in high performance organic solar cells, benchmarked against conventional indium tin oxide (ITO) electrodes, is assessed.

51 52 Author Index

Alves Perdigao Marwa, 31 Luis, 26 Horbury Anonye Michael, 9 Blessing, 22 Huang Atalay Qi, 32 Selcuk, 19 Awasthi Irshad Shalini, 27 Zehra, 6

Balasundaram Jacomin Manikandan, 15 Anne-Claire, 35 Bhandare Kempa Amol, 3 Dominik, 13 Darlington Kubes Alexander, 5 Martin, 18 Davis Kumar James, 21 Swati, 44 Dimitri Lee Federica, 17 Jaemin, 12 Edwards Lucas Ingrid, 16 Emma, 38 Lucio Fonseka Anthony, 39 Aruni, 28 Lyman Joseph, 40 Gotze¨ Kathrin, 2 Onat Gharanei Berk, 41 Seley, 29 Ghosal Palani Ritika, 14 Saravanan, 34 Gupta Patel Vinay Kumar, 30 Anisha, 23 Patrick Harrison Christopher, 42 Jonathan, 37 Poluektov Hassan Mikhail, 8

53 Pomiro Maryam, 46 Fernando, 43 Sidda Pracharova John, 47 Jitka, 36 Siew Cynthia, 7 Rana Spiga Anirudh, 4 Jenny, 48 Read Staniforth Tania, 45 Michael, 49 Roesner Stefan, 33 Townsend Eleanor, 50 Shantikumar Saran, 20 Varagnolo Sharifzadeh Silvia, 51