Computing and Information Systems - Research Publications
Permanent URI for this collection
Search Results
1 - 5 of 5
-
ItemMachine learning for classification of hypertension subtypes using multi-omics: A multi-centre, retrospective, data-driven study(ELSEVIER, 2022-10)BACKGROUND: Arterial hypertension is a major cardiovascular risk factor. Identification of secondary hypertension in its various forms is key to preventing and targeting treatment of cardiovascular complications. Simplified diagnostic tests are urgently required to distinguish primary and secondary hypertension to address the current underdiagnosis of the latter. METHODS: This study uses Machine Learning (ML) to classify subtypes of endocrine hypertension (EHT) in a large cohort of hypertensive patients using multidimensional omics analysis of plasma and urine samples. We measured 409 multi-omics (MOmics) features including plasma miRNAs (PmiRNA: 173), plasma catechol O-methylated metabolites (PMetas: 4), plasma steroids (PSteroids: 16), urinary steroid metabolites (USteroids: 27), and plasma small metabolites (PSmallMB: 189) in primary hypertension (PHT) patients, EHT patients with either primary aldosteronism (PA), pheochromocytoma/functional paraganglioma (PPGL) or Cushing syndrome (CS) and normotensive volunteers (NV). Biomarker discovery involved selection of disease combination, outlier handling, feature reduction, 8 ML classifiers, class balancing and consideration of different age- and sex-based scenarios. Classifications were evaluated using balanced accuracy, sensitivity, specificity, AUC, F1, and Kappa score. FINDINGS: Complete clinical and biological datasets were generated from 307 subjects (PA=113, PPGL=88, CS=41 and PHT=112). The random forest classifier provided ∼92% balanced accuracy (∼11% improvement on the best mono-omics classifier), with 96% specificity and 0.95 AUC to distinguish one of the four conditions in multi-class ALL-ALL comparisons (PPGL vs PA vs CS vs PHT) on an unseen test set, using 57 MOmics features. For discrimination of EHT (PA + PPGL + CS) vs PHT, the simple logistic classifier achieved 0.96 AUC with 90% sensitivity, and ∼86% specificity, using 37 MOmics features. One PmiRNA (hsa-miR-15a-5p) and two PSmallMB (C9 and PC ae C38:1) features were found to be most discriminating for all disease combinations. Overall, the MOmics-based classifiers were able to provide better classification performance in comparison to mono-omics classifiers. INTERPRETATION: We have developed a ML pipeline to distinguish different EHT subtypes from PHT using multi-omics data. This innovative approach to stratification is an advancement towards the development of a diagnostic tool for EHT patients, significantly increasing testing throughput and accelerating administration of appropriate treatment. FUNDING: European Union's Horizon 2020 Research and Innovation Programme under Grant Agreement No. 633983, Clinical Research Priority Program of the University of Zurich for the CRPP HYRENE (to Z.E. and F.B.), and Deutsche Forschungsgemeinschaft (CRC/Transregio 205/1).
-
ItemDevelopment of grid frameworks for clinical trials and epidemiological studies(IOS Press, 2006)E-Health initiatives such as electronic clinical trials and epidemiological studies require access to and usage of a range of both clinical and other data sets. Such data sets are typically only available over many heterogeneous domains where a plethora of often legacy based or in-house/bespoke IT solutions exist. Considerable efforts and investments are being made across the UK to upgrade the IT infrastructures across the National Health Service (NHS) such as the National Program for IT in the NHS (NPFIT) [1]. However, it is the case that currently independent and largely non-interoperable IT solutions exist across hospitals, trusts, disease registries and GP practices – this includes security as well as more general compute and data infrastructures. Grid technology allows issues of distribution and heterogeneity to be overcome, however the clinical trials domain places special demands on security and data which hitherto the Grid community have not satisfactorily addressed. These challenges are often common across many studies and trials hence the development of a re-usable framework for creation and subsequent management of such infrastructures is highly desirable. In this paper we present the challenges in developing such a framework and outline initial scenarios and prototypes developed within the MRC funded Virtual Organisations for Trials and Epidemiological Studies (VOTES) project [2].
-
ItemSingle sign-on and authorization for dynamic virtual organizations(Springer, 2006)The vision of the Grid is to support the dynamic establishment and subsequent management of virtual organizations (VO). To achieve this presents many challenges for the Grid community with perhaps the greatest one being security. Whilst Public Key Infrastructures (PKI) provide a form of single sign-on through recognition of trusted certification authorities, they have numerous limitations. The Internet2 Shibboleth architecture and protocols provide an enabling technology overcoming some of the issues with PKIs however Shibboleth too suffers from various limitations that make its application for dynamic VO establishment and management difficult. In this paper we explore the limitations of PKIs and Shibboleth and present an infrastructure that incorporates single sign-on with advanced authorization of federated security infrastructures and yet is seamless and targeted to the needs of end users. We explore this infrastructure through an educational case study at the National e-Science Centre (NeSC) at the University of Glasgow and Edinburgh.
-
ItemThe brain monitoring with information technology (BrainIT) collaborative network: EC feasibility study results(Springer, 2009)BACKGROUND: The BrainIT group works collaboratively on developing standards for collection and analyses of data from brain injured patients towards providing a more efficient infrastructure for assessing new health care technology. EC funding supported meetings over a year to discuss and define a core dataset to be collected with IT based methods from patients with traumatic brain injury. We now report on the results of a follow-up period of funding to test the feasibility for collection of the core dataset with IT based methods. METHODS: Over a three year period, data collection client and web-server based tools were developed and core data (grouped into 9 categories) were collected from 200 head-injured patients by local nursing staff. Data were uploaded by the BrainIT web and random samples of received data were selected automatically by computer for validation by data validation (DV) research nurse staff against gold standard sources held in the local centre. Validated data were compared with original data sent and percentage error rates calculated by data category. Feasibility was assessed in terms of the amount of missing data, accuracy of data collected and limitations reported by users of the IT methods. FINDINGS: Thirteen percent of data files required cleaning. Thirty “one-off” demographic and clinical data elements had significant amounts of missing data (> 15%). Validation nurses conducted 19,461 comparisons between uploaded database data with local data sources and error rates were generally less than or equal to 6%, the exception being the surgery data class where an unacceptably high error rate was found. Nearly 10,000 therapies were successfully recorded with start-times but approximately a third had inaccurate or missing end times which limits analyses assessing duration of therapy. Over 40,000 events and procedures were recorded but events with long durations (such as transfers) were more likely to have “end-times” missed. CONCLUSIONS: The BrainIT core dataset is a rich dataset for hypothesis generation and post-hoc analyses provided studies avoid known limitations in the dataset. Limitations in the current IT based data collection tools have been identified and have been addressed. Future academic led multi-centre data collection projects must decrease validation costs and likely will require more direct electronic access to hospital based clinical data sources for both validation purposes and for reducing the research nurse time needed for double data entry. This type of infrastructure will foster remote monitoring of patient management and protocol adherence in future trials of patient management and monitoring.
-
ItemSupporting grid-based clinical trials in Scotland(Sage, 2008)A computational infrastructure to underpin complex clinical trials and medical population studies is highly desirable. This should allow access to a range of distributed clinical data sets; support the effi cient processing and analysis of the data obtained; have security at its heart; and ensure that authorized individuals are able to see privileged data and no more. Each clinical trial has its own requirements on data sets and how they are used; hence a reusable and fl exible framework offers many advantages. The MRC funded Virtual Organisations for Trials and Epidemiological Studies (VOTES) is a collaborative project involving several UK universities specifi cally to explore this space. This article presents the experiences of developing the Scottish component of this nationwide infrastructure, by the National e-Science Centre (NeSC) based at the University of Glasgow, and the issues inherent in accessing and using the clinical data sets in a fl exible, dynamic and secure manner.