Thomas Young Centre

The extracellular matrix (ECM) defines the biomechanical and biochemical microenvironment of tissues, directing cell behaviour and phenotype. In the ovary, ECM must dynamically remodel in each cycle under hormonal regulation to control follicle development and produce fertilizable oocytes. Dysregulation of this process may result in aberrant formation of ECM as seen in polycystic ovary syndrome (PCOS) whose pathology includes fibrosis of the ovary and which is a major cause of infertility. PCOS is characterised by hyperandrogenism and, here, we investigate the impact of androgens on fibrosis, cell-ECM interactions and mechanosensing. We report an altered network of gene expression related to the genesis of fibrosis. Preantral follicles from C57BL/6 mice (14-15 days postpartum) were stimulated with dihydrotestosterone (DHT, 10nM) in 24/72 hours culture. Expression of fibrosis-associated genes (Eln; Ctgf; Acta2; Plod2; Hpse) significantly increased with androgen (72h), as did TGF-β signalling (Tgfb1; Tgfb3). We show a direct connection between androgen and mechanosensing within the ovary, with androgen upregulating the mechanosensitive Hippo pathway (Yap1; Lats1; Lats2; Stk3; Stk4; Frmd6) and downstream targets (Ctgf; Axl; Cyr61). Our results highlight hyperandrogenism as a probable driver of the fibrosis in the polycystic ovary, and emphasise the importance of ECM regulation in follicle development and fertility.

Abstract Training machine learning interatomic potentials (MLIPs) on total energies of molecular clusters using differential or transfer learning is becoming a popular route to extend the accuracy of correlated wave-function theory to condensed phases. A key challenge, however, lies in validation, as reference observables in finite-temperature ensembles are not available at the reference level. Here, we construct synthetic reference data from pretrained MLIPs and evaluate the generalizability of cluster-trained models on ice-Ih, considering scenarios where both energies and forces and where only energies are available for training. We study the accuracy and data-efficiency of differential, single-fidelity transfer, and multi-fidelity transfer learning against ground-truth thermodynamic observables. We find that transferring accuracy from clusters to bulk requires regularization, which is best achieved through multi-fidelity transfer learning when training on both energies and forces. By contrast, training only on energies introduces artefacts: stable trajectories and low energy errors conceal large force errors, leading to inaccurate microscopic observables. More broadly, we show that accurate reproduction of microscopic structure correlates strongly with low force errors but only weakly with energy errors, whereas global properties such as energies and densities correlate with low energy errors. This
highlights the need to incorporate forces during training or to apply careful validation before production. Our results highlight the promise and pitfalls of cluster-trained MLIPs for condensed phases and provide guidelines for developing - and critically, validating - robust and data-efficient MLIPs.

DEAD-box proteins such as Dhh1 play essential roles in RNA metabolism and the formation of biomolecular condensates, with a modular architecture comprising folded domains and disordered regions. To elucidate how this architecture shapes conformational dynamics in solution, we combined solution scattering experiments and multi-scale simulations on core and full-length constructs. Enhanced-sampling simulations captured a dynamic ensemble of core conformations stabilized by transient interdomain contacts that underpin functional regulation. Coarse-grained modelling revealed that disordered tails behave as independent modules exerting minimal influence on core dynamics. This integrated approach reveals a modular organization balancing structural heterogeneity and functional specificity, providing a framework for studying DEAD-box proteins in phase separation.

This article rethinks meritocratic ideology as practical knowledge that transforms through biographies of social and geographical mobility. Drawing on 37 interviews with Black and Muslim Italians living in Britain or returned to Italy, the article shows that meritocracy is rarely invoked as a coherent ideology but works as practical, embodied commonsense about the world order, with Britain leading a hierarchy of European societies. The article explores three dimensions of meritocratic commonsense and racialised minorities' double-consciousness (Du Bois). First, 'meritocratic Britain' is not simply a neoliberal narrative, but draws from postcolonial, intergenerational histories of family migration that include desires for equality and security. Second, participants' encounters with British racism do not necessarily challenge beliefs in meritocratic Britain, as being racialised as 'foreigners' in Italy leaves deeper scars on their sense of identity, belonging and recognition. Third, meritocratic Britain can lose emotional resonance when participants feel desires for connectedness and home that are not satisfied by occupational and educational mobility. By centring racialised minorities' double-consciousness, practical knowledge and struggles for recognition, the article highlights the limitations of false consciousness, misinformation and psychological compensation as explanations for meritocratic belief. Moreover, it unravels how meritocratic narratives transform across life stages.

Understanding how solvent molecules participate in chemical reaction mechanisms remains a central challenge in molecular simulations. Here, we investigate the mechanism and kinetics of glucose mutarotation in aqueous solution using a combined well-tempered metadynamics and mean force integration approach, within the framework of canonical transition state theory. We compute free energy landscapes and kinetic rate coefficients for the α → β mutarotation in pure water, as well as in systems representative of water/methanol and water/acetone mixtures. Our simulations indicate that ring opening and closure occur via concerted, solvent-assisted pathways, with ring opening identified as the rate-determining step. The temperature dependence of the predicted kinetic constant is quantitatively consistent with experimental data, while the reactivity modulation by organic co-solvents is captured qualitatively. Overall, the results provide molecular-level insight into solvent-mediated carbohydrate chemistry, demonstrating that the solvent participates in all key reactive steps and highlighting how enhanced sampling techniques can yield mechanistic and kinetic information in explicitly solvated reactive systems.

Detailed balance, the correct thermalization of electronic state populations, is an essential and elusive property of quantum-classical non-adiabatic dynamics methods. While some methods can reproduce detailed balance through physically well-motivated algorithmic adaptations, or by construction of a conserved Hamiltonian function, the physical mechanism leading to detailed balance is not understood from first principles. Coupled trajectory mixed quantum-classical (CTMQC) dynamics may provide some insight into the question, as it can be derived from first principles in the exact factorization theorem of full quantum mechanics. Although we find that the current conventional flavor of CTMQC, which conserves energy across the ensemble of trajectories (known as CTMQC-E), fails to reproduce detailed balance as in Ehrenfest dynamics, we show that a similar variant, where total energy is conserved on each trajectory independently, provides a major improvement over Ehrenfest with respect to detailed balance. Moreover, we show that the theory achieves convergence of the mean electronic potential energy with the number of energy levels that successively increase in energy. This new variant is shown to, by simulations on the Tully models and double arch model, retain a good description electronic populations and coherence compared to exact quantum dynamics. We explain the thermalization mechanism through the additional terms that distinguish CTMQC from Ehrenfest dynamics. We show that the improvement can be explained via geometric contributions to the nuclear force, resulting from the quantum momentum, which act to oppose motion when electrons decohere upward in energy and act to enhance motion otherwise, somewhat emulating the mechanism of frustrated hops. These results have considerable implications for the applicability of CTMQC to condensed phase simulations.

Pnictogen chalcohalides (MChX) represent an emerging class of nontoxic photovoltaic absorbers, valued for their favorable synthesis conditions and optoelectronic properties. Despite their proposed defect tolerance, stemming from the antibonding nature of their valence and conduction bands, their experimentally reported power conversion efficiencies remain below 10%, far from the ideal Shockley-Queisser limit of 30%. Using advanced first-principles simulation methods, we uncover a complex point-defect landscape in MChX, exemplified by BiSeI. Previously overlooked chalcogen vacancies are identified as critical nonradiative charge-recombination centers, which exist in high concentrations and, although they exhibit modest capture coefficients, can reduce the maximum power conversion efficiency down to 24%. We argue that such detrimental effects can be mitigated by cation-poor synthesis conditions and strategic anion substitutions. This study not only identifies efficiency-limiting factors in MChX but also provides a roadmap for their improvement, paving the way for next-generation solution-processed chalcogenide photovoltaics.

While there have been many developments in computational probes of both strongly-correlated molecular systems and machine-learning accelerated molecular dynamics, there remains a significant gap in capabilities in simulating accurate non-local electronic structure over timescales on which atoms move. We develop an approach to bridge these fields with a practical interpolation scheme for the correlated many-electron state through the space of atomic configurations, whilst avoiding the exponential complexity of these underlying electronic states. With a small number of accurate correlated wave functions as a training set, we demonstrate provable convergence to near-exact potential energy surfaces for subsequent dynamics with propagation of a valid many-body wave function and inference of its variational energy whilst retaining a mean-field computational scaling. This represents a profoundly different paradigm to the direct interpolation of potential energy surfaces in established machine-learning approaches. We combine this with modern electronic structure approaches to systematically resolve molecular dynamics trajectories and converge thermodynamic quantities with a high-throughput of several million interpolated wave functions with explicit validation of their accuracy from only a few numerically exact quantum chemical calculations. We also highlight the comparison to traditional machine-learned potentials or dynamics on mean-field surfaces.

CO2 capture using carbon-based materials, particularly graphene and graphene-like materials, is a promising strategy to deal with CO2 emissions. However, significant gaps remain in our understanding of the molecular-level interaction between CO2 molecules and graphene, particularly in terms of chemical bonding and electron transfer. In this work, we employ random structure search and density functional theory to understand the adsorption of CO2 molecules on Ca-, Sr-, Na-, K-, and Ti-decorated graphene surfaces. Compared to the pristine material, we observe enhanced CO2 adsorption on the decorated graphene surfaces. Particularly on group 2 metals and titanium-decorated graphene, CO2 can be strongly chemisorbed as a bent CO2 anion or as an oxalate, depending on the number of CO2 molecules. Electronic structure analysis reveals that the adsorption mechanism involves a charge transfer from the metal adatom to the adsorbed CO2. Overall, this study suggests that reducing CO2 to oxalate on group 2 metals and titanium-decorated graphene surfaces is a potential strategy for CO2 storage. Published by the American Physical Society 2025

Quantum Monte Carlo (QMC) methods represent a powerful family of computational techniques for tackling complex quantum many-body problems and performing calculations of stationary state properties. QMC is among the most accurate and powerful approaches to the study of electronic structure, but its application is often hindered by a steep learning curve; hence it is rarely addressed in undergraduate and postgraduate classes. This tutorial is a step toward filling this gap. We offer an introduction to the diffusion Monte Carlo (DMC) method, which aims to solve the imaginary time Schrödinger equation through stochastic sampling of the configuration space. Starting from the theoretical foundations, the discussion leads naturally to the formulation of a step-by-step algorithm. To illustrate how the method works in simplified scenarios, examples such as the harmonic oscillator and the hydrogen atom are provided. The discussion extends to the fixed-node approximation, a crucial approach for addressing the fermionic signproblem in multi-electron systems. In particular, we examine the influence of trial wave function nodal surfaces on the accuracy of DMC energy by evaluating results from a non-interacting two-fermion system. Extending the method to excited states is feasible in principle, but some additional considerations are needed, supported by practical insights. By addressing the fundamental concepts from a hands-on perspective, we hope this tutorial will serve as a valuable guide for researchers and students approaching DMC for the first time.