A5031010679- Statistical Methods and Bayesian Inference
- Inflammatory Bowel Disease
- Statistical Methods and Inference
- Gaussian Processes and Bayesian Inference
- Epigenetics and DNA Methylation
- Chronic Disease Management Strategies
- Microscopic Colitis
- Dementia and Cognitive Impairment Research
- Genetic Associations and Epidemiology
- Bayesian Modeling and Causal Inference
- GDF15 and Related Biomarkers
- Biomarkers in Disease Mechanisms
- S100 Proteins and Annexins
- Eosinophilic Esophagitis
- Soil Geostatistics and Mapping
- Helicobacter pylori-related gastroenterology studies
- IL-33, ST2, and ILC Pathways
- Celiac Disease Research and Management
- Acute Myocardial Infarction Research
- Cardiac Imaging and Diagnostics
- Control Systems and Identification
- Cardiac electrophysiology and arrhythmias
- Cardiovascular Function and Risk Factors
Institute of Genetics and Cancer
2021-2025
University of Edinburgh
2021-2025
Edinburgh Cancer Research
2025
University of Warwick
2018-2020
Type 2 diabetes mellitus (T2D) presents a major health and economic burden that could be alleviated with improved early prediction intervention. While standard risk factors have shown good predictive performance, we show the use of blood-based DNA methylation information leads to significant improvement in 10-year T2D incidence risk. Previous studies been largely constrained by linear assumptions, cytosine–guanine pairs one-at-a-time binary outcomes. We present flexible approach (via an R...
Despite poor cardiovascular outcomes, there are no dedicated, validated risk stratification tools to guide investigation or treatment in type 2 myocardial infarction. The goal of this study was derive and validate a tool for the prediction death future infarction patients with T2-risk score developed prospective multicenter cohort consecutive Cox proportional hazards models were constructed primary outcome at 1 year using variables selected priori based on clinical importance. Discrimination...
Abstract Background Despite major advances in our understanding of the molecular underpinning IBD, we remain challenged deciphering observed disease heterogeneity. Attempts to characterise longitudinal behaviour have either been restricted symptom-based profiling - as exemplified IBSEN cohorts or reliance on stochastic endpoints “progression” such hospitalisation, surgery and fibrosis. Here present a novel method examine IBD by modelling inflammatory patterns two large, well-characterised...
The progressive nature of Crohn's disease is highly variable and hard to predict. In addition, symptoms correlate poorly with mucosal inflammation. There therefore an urgent need better characterize the heterogeneity trajectories in by utilizing objective markers We aimed understand this clustering patients similar longitudinal fecal calprotectin profiles.We performed a retrospective cohort study at Edinburgh IBD Unit, tertiary referral center, used latent class mixed models cluster subjects...
Abstract Type 2 diabetes mellitus (T2D) presents a major health and economic burden that could be alleviated with improved early prediction intervention. While standard risk factors have shown good predictive performance, we show the use of blood-based DNA methylation information leads to significant improvement in 10-year T2D incidence risk. Previous studies been largely constrained by linear assumptions, CpGs one-at-a-time, binary outcomes. We present flexible approach (via an R package,...
Abstract Background Crohn’s disease (CD) and ulcerative colitis (UC) are highly heterogeneous, dynamic unpredictable, with a marked disconnect between symptoms intestinal inflammation. Attempts to classify inflammatory bowel (IBD) subphenotypes inform clinical decision making have been limited. We aimed describe the latent heterogeneity by modelling routinely collected faecal calprotectin (FC) CRP data, describing longitudinal patterns in IBD. Methods In this retrospective study, we analysed...
Gaussian processes are valuable tools for non-parametric modelling, where typically an assumption of stationarity is employed. While removing this can improve prediction, fitting such models challenging. In work, hierarchical constructed based on Markov random fields with stochastic spatially varying parameters. Importantly, allows non-stationarity while also addressing the computational burden through a sparse banded representation precision matrix. setting, efficient chain Monte Carlo...
Abstract Background Faecal calprotectin (FCAL) is routinely used to monitor mucosal inflammation, and high FCAL values are associated with poor outcomes in IBD. Having previously found clusters characterised by distinct profiles a Crohn’s disease (CD) cohort of 356 subjects, we aimed extend this approach all IBD patients treated the Edinburgh Unit. Methods Subjects were required have diagnosis IBD, three measurements within ten years between January 2005 September 2019, be All was measured...
Background Blood-based biomarkers of brain health could provide a cost-effective contribution to detecting individuals at risk dementia. Epigenetic scores (EpiScores) for blood protein levels have previously associated with several disease outcomes and measures health, however this has typically been limited single EpiScore analyse. Results Using 84 EpiScores as candidate biomarkers, associations general cognitive function (both cross-sectionally longitudinally) were tested in three...
A bstract Background and Aims The progressive nature of Crohn’s disease is highly variable hard to predict. In addition, symptoms correlate poorly with mucosal inflammation. There therefore an urgent need better characterise the heterogeneity trajectories in by utilising objective markers We aimed understand this clustering patients similar longitudinal faecal calprotectin profiles. Methods Latent class mixed models were used model within five years diagnosis cluster subjects. Information...
Gaussian processes (GPs) are frequently used in machine learning and statistics to construct powerful models. However, when employing GPs practice, important considerations must be made, regarding the high computational burden, approximation of posterior, choice covariance function inference its hyperparmeters. To address these issues, Hensman et al. (2015) combine variationally sparse with Markov chain Monte Carlo (MCMC) derive a scalable, flexible general framework for GP Nevertheless,...