Imperial publications
Results
- Showing results for:
- Reset all filters
Search results
-
Journal articleTam KMM, Brown NC, Bronstein M, et al., 2026,
Learning constrained static equilibrium for thrust network inverse form-finding via physics-informed geometric deep learning on CW complexes
, Advanced Engineering Informatics, Vol: 70, ISSN: 1474-0346This work integrates Geometric Deep Learning (GDL) with physics-informed modelling to approximate solutions to a constrained, ill-conditioned, and nonlinear inverse shell form-finding problem across diverse geometries and patterns. Given a target geometry and pattern design as meshes, the proposed neural framework predicts a funicular shell—defined by edge forces and vertex positions—that satisfies static equilibrium and closely matches the target form. Three main contributions are introduced: (1) a relaxed, numerically stable physics objective using efficient differentiable graph operators to mitigate the ill-conditioning of the nonlinear problem; (2) a stochastic augmentation strategy that enriches training with geometries of varying funicular feasibility, enhancing generalisation to infeasible inputs; and (3) a hierarchical GDL architecture that learns directly from irregular n -gon surface meshes, modelled as cell complexes to incorporate vertex, edge, and face features in both inputs and outputs. This approach eliminates the need for simplification of graph datastructures common in existing methods, improving mesh modelling versatility. Extensive studies examine the numerical stability of physics formulations, robustness for out-of-distribution designs, and the expressivity of the GDL architecture. While focused on a specific inverse form-finding task, this work offers general insights into addressing ill-posed inverse problems, showing how physics-based learning can support optimisation of mesh-based architectural structures under variable connectivity and design constraints.
-
Journal articleHo V, Baker J, Fenwick P, et al., 2026,
Single cell microfluidic quantification of miRNA-21 and miRNA-34a reveals miRNA interactions in small airway epithelial cells and fibroblasts from COPD patients
, American Journal of Physiology - Lung Cellular and Molecular Physiology, ISSN: 1040-0605Rationale: MicroRNA-21 and microRNA-34a are implicated in chronic obstructive pulmonary disease (COPD) pathogenesis, but their cell-specific expression patterns and interactions within individual airway cells remain unexplored.Objective: To develop a single cell microfluidic platform for dual, amplification-free detection of miR-21-5p and miR-34a-5p in primary small airway cells from COPD patients.Methods: Small airway epithelial cells (SAEC) and fibroblasts (SAF) were isolated from COPD patients and non-smokers (n = 6–8 per group). A microfluidic chip with dual miRNA sandwich hybridisation assays was used to quantify miR-21-5p and miR-34a-5p in single cells. Expression of miRNAs and their target genes was evaluated under oxidative stress using qPCR and Western blotting.Main Results: Single cell analysis revealed significantly higher miR-21-5p and miR-34a-5p expression in COPD-derived cells compared to controls. MiR-21 exhibited greater variability than miR-34a, and their positive correlation in control cells was disrupted in COPD. Oxidative stress elevated miR-21 and miR-34a while reducing expression of miR-21 targets and increasing senescence markers (p21Cip1/Waf1, p16INK4a). MiR-21 antagomir restored expression of suppressed targets in both cell types.Conclusions: Our novel single cell microfluidic platform enables precise, simultaneous detection of miR-21 and miR-34a in single small airway cells. This allows the interrelationship between the miRNAs to be assessed within the same cell. MiR-21 and miR-34a represent promising therapeutic targets for restoring gene regulatory balance in COPD.
-
Journal articleHamilton MS, Derks I, Kaforou M, et al., 2026,
Identifying correlates of risk for future tuberculosis disease progression in South African children (intrepid): a protocol paper
, BMJ Open, ISSN: 2044-6055IntroductionYoung children and children living with HIV are at high risk of progressing to tuberculosis (TB) disease following Mycobacterium tuberculosis (Mtb) exposure and infection, and also of developing severe forms of disease and TB-related mortality. Identifying children who have very early (sub-clinical) TB disease, prior to progression to clinically apparent TB, would mean that TB preventive treatment (TPT) could be more efficiently targeted to this group. Identifying biomarker changes on drug therapy in children with Mtb infection or very early disease could pave the way for the development of tests that can identify which children have viable bacilli and are therefore at increased risk of disease progression. Methods and analysisThe INTREPID study will utilize already collected samples taken from well-phenotyped paediatric cohorts in three clinical studies conducted in South Africa in children <5 years, including a drug-resistant TPT trial (TB-CHAMP), an observational household contact study (IGRA studies), and a prospective diagnostic study (Umoya), all conducted in a setting with a high burden of TB and HIV. We will employ transcriptomic, proteomic, metabolomic, and serology approaches to analyse changes in host blood profiles at every stage along the TB continuum, from Mtb exposure to disease and from children treated for Mtb infection and early TB disease, as well as targeted Mtb antibody analysis. Data on viral co-infections and relevant clinical and epidemiological parameters will be integrated and evaluated to identify the optimal biosignatures that can predict future progression to clinically overt disease in children below 5 years of age, including those living with HIV. Ethics and disseminationThe study protocol received ethical approval from the Stellenbosch University Health Research Ethics Committee (N23/03/025). The study findings will be disseminated through peer-reviewed publications, scientific conferences, and formal presentations to
-
Journal articleWang D, Hadad N, Moss S, et al., 2026,
Assessment of mosaic loss of chromosome Y in pulmonary fibrosis reveals limited association with susceptibility or disease severity.
, BMJ Open Respir Res, Vol: 13BACKGROUND: Pulmonary fibrosis (PF) is a rare lung disease with diverse pathogenesis and biological mechanisms. Mosaic loss of chromosome Y (mLOY) has been reported to be associated with increased risk of fibrotic diseases. However, the exact role of mLOY in the development of PF remains to be elucidated. METHODS: Copy number on chromosome Y was used to estimate mLOY comparing patients in PROFILE and gnomAD cohorts and between cases and control patients from the GE100KGP cohort. Correlation of mLOY with demographic and clinical variables was tested using patients from the PROFILE cohort. Lung single-cell transcriptomic data were analysed to assess the cell types implicated in mLOY. Mendelian randomisation was performed to examine the causal relationship between mLOY, idiopathic pulmonary fibrosis (IPF) and telomere length. RESULTS: The genetic analysis suggests that mLOY is found in PF from both case cohorts but when compared with an age matched population the effect is minimal (p=0.00316, median: 0.288 vs 0.291). mLOY is related to age (p=0.000214) and shorter telomere length (p=0.00815) rather than PF severity or progression. Single-cell analysis indicates that mLOY appears to be found primarily in immune cells. Mendelian randomisation demonstrates that mLOY is not on the causal pathway for IPF, but partial evidence supports that telomere shortening is on the causal pathway for mLOY. CONCLUSIONS: Our study confirms the existence of mLOY in PF patients, suggests that mLOY is not a major driver of IPF, and might support a triangulation model where telomere shortening leads to both IPF and mLOY.
-
Journal articleHazell L, Cooper N, 2026,
Multi-arm multi-stage randomised controlled trial of Inflammatory Signal Inhibitors (MATIS) for patients hospitalised with COVID-19 pneumonia during the UK pandemic
, BMJ Open, ISSN: 2044-6055Objectives: To determine the safety and efficacy of ruxolitinib and fostamatinib compared to standard of care (SOC) in patients requiring hospital admission for the treatment of COVID-19 pneumonia.Design: Adaptive multi-arm, multi-stage, randomised, open label trial (3-arm, 2-stage)Setting: Five hospitals in England between October 2020 and September 2022.Participants: Hospitalised patients (≥18 years) with COVID-19 pneumonia defined by a modified World Health Organisation (WHO) COVID-19 severity grade of 3 or 4. Interventions: Participants were randomly assigned 1:1:1 to receive ruxolitinib (10mg bd for 7 days then 5mg bd for 7 days), fostamatinib (150mg bd for 7 days then 100mg bd for 7 days) or SOC.Main outcome measures: Primary outcome was development of severe COVID-19 pneumonia (modified WHO severity grade ≥5) within 14 days of randomisation. Secondary outcomes included mortality, invasive and non-invasive ventilation, venous thromboembolism, duration of hospital stay, readmissions, inflammatory markers and serious adverse events (SAE). Results: At Stage 1, 181 patients were randomised, with 4 assessed as ineligible post-randomisation. Fostamatinib was stopped early for futility with 16 participants (27.6%, N=58) developing severe COVID-19 pneumonia compared to 15 (25.0%, N=60) in the SOC arm (adjusted odds ratio (aOR) compared to SOC: 1.12; 95% confidence interval (CI): 0.49 to 2.58; p=0.608). Ruxolitinib progressed to Stage 2 but the trial was stopped early due to slow recruitment. At the final analysis, 10 participants (16.1%, N=62) developed severe COVID-19 pneumonia in the ruxolitinib arm compared to 15 (24.6%, N=61) in the SOC arm (aOR: 0.63; 95% CI: 0.25 to 1.57; p=0.161). Four (7.4%) participants in the fostamatinib arm, none in the ruxolitinib arm and three (5.5%) in the SOC arm died within 14 days of randomisation. Infections were the most frequently reported serious adverse event and were numerically higher in the fostamatinib (10, 17.2%) an
-
Journal articleHay CD, Mahutanattan SM, Pilkington CP, et al., 2026,
Affordable, cleanroom-free millifluidic production of targeted lipid nanocarriers via additive manufacturing.
, Lab Chip, Vol: 26, Pages: 635-649Lipid nanocarriers utilise the self-assembly of amphiphilic molecules to generate particle formulations capable of drug encapsulation and dynamic interactions with user-defined cell types, enabling applications within targeted therapeutic delivery. This offers increased bioavailability, stability, and reduced off-target effects, with the promise of application to numerous cell types and consequently, diseases. Here, we have developed a highly accessible, cleanroom-free method for the fabrication of poly(methyl methacrylate) millifluidic vertical flow focusing (VFF) devices via laser cutting, multilayered solvent and heat-assisted bonding. We demonstrate that these can be used for one-step production of targeted lipid nanocarriers via the production of cardiomyocyte-targeting vesicle nanoparticles loaded with the hydrophobic drug menadione. We characterise vesicle size using dynamic light scattering (DLS) and cryogenic transmission electron microscopy (cryo-TEM), whilst also probing the membrane viscosity of vesicles produced via flow-focusing for the first time using molecular rotors. Finally, we apply cardiomyocyte-targeting, menadione-loaded vesicles to H9C2 tissue culture demonstrating significant inhibition of cell viability via targeted delivery, showcasing the potential of our device to produce formulations for therapeutic delivery. As a flow-based method, VFF can facilitate rapid formulation investigation and produce large sample volumes for cell-based validation studies, whilst avoiding inter-batch sample variation. Furthermore, the accessible nature of this VFF approach will help to democratise millifluidics, facilitating the wider adoption of flow-based production methods to develop nanomedical formulations.
-
Journal articleStewart I, John A, Bin L, et al., 2026,
Residual lung abnormality following COVID-19 hospitalisation is characterised by biomarkers of epithelial injury
, EBioMedicine, Vol: 124, ISSN: 2352-3964Some survivors of acute COVID-19 infection have long-term symptoms that could suggest ongoing lung impairment. Searches performed in MEDLINE and Embase for SARS-COV-2 studies with radiological lung follow-up estimated that 50% of participants had inflammatory patterns and 29% had fibrotic patterns at a median of 3 months post infection. Analysis of the UK nationwide Post-hospitalisation COVID-19 Study at 5-months follow-up suggested that up to 11% of people discharged from hospital following COVID-19 infection were at-risk of radiological residual lung abnormalities, such as ground glass opacity and reticulation. In people with pulmonary fibrosis, these radiological patterns are often consistent with persistent epithelial lung injury. Biomarker studies have identified associations with COVID-19 severity, however there are few studies that explore the relationship between biomarkers of epithelial injury and parenchymal lung abnormalities post-hospitalisation.
-
Journal articleFeng Q, Manousou P, IzziEngbeaya CN, et al., 2026,
Unveiling the burden of steatotic liver disease: mortality risks by subtype and fibrosis stage in a nationwide cohort
, Liver International, Vol: 46, ISSN: 1478-3223Background and AimsWe investigated the associations between SLD, fibrosis stage, and all-cause and cause-specific mortality, with a focus on SLD subtypes.MethodsWe analysed 486 156 UK Biobank participants. SLD cases were identified using fatty liver index ≥ 60. Causes of death were confirmed via death registries. Multivariable Cox models estimated associations between SLD, SLD subtypes, FIB4 score and mortality outcomes, including all-cause mortality, mortality from liver-related diseases, cardiovascular disease (CVD) and extrahepatic cancers.ResultsSLD was identified in 178 336 participants (36.7%): 73.5% with MASLD, 19.0% with MetALD and 6.4% with ALD. Over a median follow-up of 13.8 years, 20 766 (11.6%) deaths occurred among people with SLD and 21 754 among those without (307 820; 7.1%), suggesting a higher mortality rate in SLD than in non-SLD (8.78 vs. 5.25/1000 person-years). All SLD subtypes were associated with higher all-cause mortality: MASLD (HR (95% CI): 1.32 (1.29–1.35)), MetALD (1.16 (1.12–1.20)) and ALD (1.36 (1.29–1.44)). Excess mortality was primarily driven by extrahepatic cancer (42.5%) and cardiovascular disease (24.2%), while liver-related deaths were concentrated among those with ALD and fibrosis. A strong dose–response relationship was observed between FIB4 stratification and mortality, particularly for liver-related deaths. These associations were independent of socioeconomic status, lifestyle and cardiometabolic risk factors.ConclusionSLD is independently associated with increased all-cause and cause-specific mortality, with substantial variation across subtypes and fibrosis severity. Extrahepatic cancer and cardiovascular disease are the leading contributors to excess mortality. These findings underscore the need for integrated care strategies targeting metabolic risk, fibrosis progression and cancer prevention in the SLD population.
-
Journal articleMuraro P, Kazmi M, De Matteis E, et al., 2026,
Real-world effectiveness of autologous haematopoietic stem cell transplantation for MS in the UK
, Journal of Neurology, Neurosurgery and Psychiatry, Vol: 97, Pages: 146-155, ISSN: 0022-3050Background Autologous haematopoietic stem cell transplantation (AHSCT) is increasingly used as a one-off disease-modifying therapy for aggressive forms of multiple sclerosis (MS). We report real-world effectiveness of AHSCT for MS in the UK.Methods This retrospective open-label study included patients with (pw)MS treated with AHSCT between 2002 and 2023 in 14 UK centres. Outcomes included relapse-free survival (RFS), MRI activity-free survival (MFS), progression-free survival (PFS) and no evidence of disease activity (NEDA-3). We assessed 6-month confirmed Expanded Disability Status Scale (EDSS) score progression or improvement compared with pre-treatment. Treatment-related mortality (TRM) was defined as death from any cause within 100 days post-autologous graft reinfusion.Results 364 pwMS were included (median age 40 years; 58% female). Of these, 271 pwMS had adequate neurological follow-up data: 168 (62%) had relapsing-remitting MS (pwRRMS) and 103 (38%) had progressive MS (pwPMS). Median disease duration from symptom onset was 10 years (IQR 6–14), EDSS 6 (IQR 4.0–6.5) and follow-up from AHSCT 46 months. At 2 and 5 years from AHSCT, RFS was 94.6% and 88.6%; MFS 93.1% and 80.1%; PFS 83.5% and 62.4%; NEDA-3 72.3% and 46.2%. pwRRMS had significantly higher rates of PFS (p=0.007) and NEDA-3 (p=0.001) than pwPMS. RRMS was a predictor of EDSS improvement, whose prevalence was 24.2% at 2 years and 20.4% at 5 years. TRM was 1.4% (n=5/364).Conclusions In this cohort with high EDSS at baseline and including pwPMS, AHSCT led to durable remission of inflammatory activity and stabilisation or improvement of neurological disability, particularly in pwRRMS.
-
Journal articleGray JL, Xiao Z, Rogga VV, et al., 2026,
From Serendipity to Strategy: Rationalizing Molecular Glue Discovery and Proximity-Induced Pharmacology through Chemical Biology.
, J Am Chem SocMolecular glues represent a new paradigm in drug discovery, stabilizing novel protein-protein interactions between two proteins to elicit targeted cellular outcomes. Historically discovered through serendipity, molecular glue identification is becoming increasingly systematic. In this perspective, we discuss the current advances and challenges of this field, highlighting the transition toward rational discovery through the convergence of four complementary approaches. Innovations in library design and screening platforms are expanding access to glue-relevant chemical space, guided by a deeper mechanistic understanding of proximity-induced pharmacology. These efforts are further enabled by functional genomic approaches that reveal gluable interfaces. Finally, the integration of chemical and biological data through machine learning is beginning to support rational de novo glue design.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.