Denaxas Lab

Electronic Health Record phenotyping
AF Raw EHR require a substantial amount of work before they can be transformed into research-ready datasets that can be analyzed to answer clinically meaningful questions. Our lab develops computational algorithms for defining, validating and ascertaining multi-modal disease phenotypes in EHR data. Created phenotypes are stored in an open-access Data Portal.
More information: Atrial fibrillation phenotyping exemplar in PLOS ONE.
Supervised machine learning for risk stratification
AF The majority of traditional risk prediction approaches rely on regression based statistical approaches and potentially fail to take into account the richness of electronic health record data. We are developing novel supervised machine learning methods that exploit the dimensionality and complexity of the data in order to create accurate disease risk prediction models in chronic and acute settings.
Unsupervised machine learning for sub-phenotype discovery
AF There is a growing body of evidence from observational and interventional research studies suggesting that complex diseases, such as type-II diabetes, asthma and COPD, are not a single disease but are composed of distinct sub-phenotypes with different risk factor and prognostic profiles. We are developing and evaluating novel methods using data clustering algorithms to identify, describe and evaluate such sub-phenotypes.
Phenome-wide association studies
AF Phenome-wide association studies (PheWAS), enable scientists to analyze multiple phenotypes compared to a single genetic variant. Our lab is building computational tools and methods for defining and validating disease phenotypes in the UK Biobank from electronic health records across primary, secondary and tertiary care.
Dementia
AF Dementia is common condition defined as a loss of mental ability severe enough to interfere with normal activities of daily living. Working with the National Institute of Health Research Queen Square Dementia Biomedical Research Unit, our lab seeks to create novel machine learning methods for detecting and quantifying cognitive decline in primary care electronic health records.