This project aims to translate a complex transcriptomics-driven risk prediction model for tuberculosis (TB) into a tool which can be used in clinical practice.
TB, an infectious disease caused by Mycobacterium tuberculosis (Mtb) remains a leading cause of global morbidity and mortality. The World Health Authority END-TB elimination strategy (2015) identifies prevention through screening and treatment of disease-free people with Mtb infection. However, current biomarkers to identify those at risk lack sufficient predictive power for feasible/cost-effective mass screening programmes. Recent studies suggest blood transcriptomic profiling may help. However, clinical translation has been limited by practical challenges such as harmonising batch variability across studies and instability of selected features leading to poor generalisability across studies.
We have produced a machine learning (ML) model using blood RNA-Seq data that demonstrates stability and reproducibility for stratifying TB-risk across multiple independent cohorts, and outperforms demographics-based predictive models. Separately we have also developed an approach simplifying implementation of clinical decision support systems (CDSS).
In this project, the student will combine both approaches by adapting the concept ML model to utilise a curated repository of TB RNA-Seq data; and produce infrastructure and tooling to create a generalisable CDSS to prospectively risk stratify patients with TB infection, identifying those at higher risk of TB progression. This has application and relevance to global TB control strategies. In collaboration with the Institute for Precision Health (IPH) we will work with commercial partners to explore development of a nanoString-based translational pipeline for clinical utilisation. Furthermore, analytical and translational principles underpinning the infrastructure will be generalisable to other disease areas exploiting high-dimensional omics data for clinical biomarker development.
The focus of this project on prevention and risk stratification fits the remit of the Institute, with the Leicester NIHR Biomedical Research Centre (BRC) providing an ideal environment for this multi-disciplinary project. The BRC has a new focus on informatics/training and growing links with the College of Science and Engineering, while the BRC-based supervisory team have extensive experience of software engineering/application and CDSS development (Free), advanced mathematical modelling and statistics (Richardson) and TB/clinical knowledge and application in TB (Haldar).
The project offers an exciting opportunity to work on the development and application of stable and robust models for high dimensional data and the challenging problem of integrating such models with decision support systems.
Entry requirements
UK Bachelor's Degree with at least 2:1 (or overseas equivalent) in a relevant subject e.g. statistics, applied mathematics, computer science or a subject with a substantial statistical or applied mathematics component.
University of Leicester English language requirements apply where applicable.
How to Apply
To apply for this position please follow the guidance at:
https://le.ac.uk/study/research-degrees/funded-opportunities/iph
Project enquiries
Dr Matthew Richardson [Email Address Removed]
Dr Rob Free [Email Address Removed]