Telomere length and vascular disease, investigating the role of gene expression through Machine Learning.

Supervisors:

• Dr Christopher P Nelson chris.p.nelson@leicester.ac.uk
• Professor Matt Bown
• Dr Svletlana Stoma
• Professor Veryan Codd

Project Description

Telomeres and Disease:  Telomeres are protein bound tandem hexamer DNA repeats which are protective structures at the end of chromosomes. The length of the telomere sequence declines with progressive cellular division. Telomere length (TL) is therefore reflective of cellular age and is also seen to decline with increasing age in humans. In addition, TL shows wide inter-individual variation at any age throughout the life course and is a heritable trait. Originally considered a potential marker of biological ageing that could potentially explain part of the inter-individual variance in age-related disease risk, TL has been associated with an increased risk of several vascular diseases including abdominal aortic aneurysm (AAA)1 and coronary artery disease (CAD)2,3. This potentially causal relationship4-6 has an overlap in signal between AAA/CAD with TL associated loci that requires further elucidation to determine shared biology and infer biological mechanism. 

Hypothesis: We hypothesise that the link between vascular diseases and TL is at least partly driven through the regulation of gene expression. Gene expression of relevant cell types can be predicted by using genetic variants by applying machine learning and deep learning methods. The estimated gene expression can then be assessed for direct and indirect effects. 

Data sources: Gene expression can be cell specific and genome wide analyses of TL with expression are generally small7,8. We have gene expression datasets for ~1,500 Vascular Smooth Muscle Cells (VSMCs)9 and ~550 monocytes and macrophages10. There are also TL measurements available for the majority of these samples where small-scale analyses can be used to assess correlation between TL and gene expression and to perform gene set enrichment analyses to identify genes and pathways with differential expression. There will also be access to three major cohorts. 1) UK Biobank: a large population-based cohort of over 500k individuals11 where we in Leicester measured leukocyte TL establishing the largest collection of telomere lengths in the world to date12 (Project 6077). 2) GENVASC: We established the Genetics and the Vascular Health Check Study (GENVASC)13 within the NHS health check framework recruiting over 44k individuals from Leicestershire and Northampton. Longitudinal data are available from both primary and secondary care, along with blood samples from which we have attained genotyping data. 3) UKAGS, the UK Aneurysm Growth Study14, is a large set of AAA cases with genetic data collected in Leicester that is a major contributor to AAAgen15.

Methodology: Machine learning can be applied to investigate the role of gene expression in the TL-AAA relationship with aims to identify the mediation of gene transcription between genetic variants and outcome. These are known as TWAS16, where models of gene expression, estimated via SNP expression associations, are trained as functions of genetic variants, similar to polygenic scores, and then applied to genome-wide association study (GWAS) data. The use of available expression resources in Leicester can utilise TWAS to generate expression predictions in UK Biobank, UKAGS and the GENVASC cohort. This post-GWAS analysis identifies gene-trait associations, functional genes regulated by genomic variation with high interpretability, enabling mechanistic relationships to be investigated to elucidate the TL->AAA link alongside associated traits. There is also an opportunity to develop and apply Deep Learning models to genomic sequence data to predict expression17. Confirmation of findings within GENVASC will involve training in core molecular laboratory techniques including, DNA and RNA extraction, qPCR and RNAseq.

Outline: This PhD offers the opportunity for multidisciplinary training and experience, with statistical modelling and machine learning forming a large part of the analysis in years 1 and 2, with year 3 having a larger focus on laboratory work as briefly detailed in this potential plan:

• Year one - Genome-wide eQTL analyses: Telomere AAA co-expression; Causal mediation; Whole genome eQTL analyses.
• Year two - Validate TWAS and test predictions: Build TWAS to determine prediction accuracy; Apply in cohorts; Association of TWAS with traits
• Year three - Validate and replicate in GENVASC: Measure telomere length; Experimental work on selected gene expression; Replication of findings

References:

1. Atturu G, et al. Short leukocyte telomere length is associated with abdominal aortic aneurysm (AAA). Eur J Vasc Endovasc Surg 39, 559-64. (2010)
2. Brouilette, S. W. et al. Telomere length, risk of coronary heart disease, and statin treatment in the West of Scotland Primary Prevention Study: a nested case-control study. Lancet 369, 107–114 (2007).
3. Fitzpatrick, A. L. et al. Leukocyte telomere length and cardiovascular disease in the cardiovascular health study. Am. J. Epidemiol. 165, 14–21 (2007).
4. Haycock, P. et al. Association Between Telomere Length and Risk of Cancer and Non-Neoplastic Diseases: A Mendelian Randomization Study. JAMA Oncol. 3, 636-651. (2017)
5. Li, C. et al. Genome-wide Association Analysis in Humans Links Nucleotide Metabolism to Leukocyte Telomere Length. Am. J. Hum. Genet. 106, 389–404 (2020).
6. Codd, V. et al. Polygenic basis and biomedical consequences of telomere length variation. Nat. Genet. 53, 1425–1433 (2021).
7. Chang, Y. et al. Unraveling the causal genes and transcriptomic determinants of human telomere length. Nat Commun 14, 8517 (2023).
8. Andreu-Sánchez, S., et al. Genetic, parental and lifestyle factors influence telomere length. Commun Biol 5, 565 (2022). 
9. Solomon, C. U. et al. Effects of Coronary Artery Disease-Associated Variants on Vascular Smooth Muscle Cells. Circulation 101161CIRCULATIONAHA121058389 (2022).
10. Garnier, S. et al. Genome-wide haplotype analysis of cis expression quantitative trait loci in monocytes. PLoS Genet. 9, e1003240 (2013).
11. Bycroft, C. et al. The UK Biobank resource with deep phenotyping and genomic data. Nature 562, 203–209 (2018).
12. Codd, V. et al. Measurement and initial characterization of leukocyte telomere length in 474,074 participants in UK Biobank. Nature Aging 2, 170–179 (2022).
13. Samani, N.J., et al. Polygenic risk score adds to a clinical risk score in the prediction of cardiovascular disease in a clinical setting, European Heart Journal, 45, 3152–3160 (2024) 
14. Bath MF, et al. Impact of abdominal aortic aneurysm screening on quality of life. Br J Surg. 105, 203-208 (2018).
15. Roychowdhury T. et al. Genome-wide association meta-analysis identifies risk loci for abdominal aortic aneurysm and highlights PCSK9 as a therapeutic target. Nat Genet. 55, 1831-1842 (2023).
16. Mai, J. et al. Transcriptome-wide association studies: recent advances in methods, applications and available databases. Commun Biol 6, 899 (2023). 
17. Barbadilla-Martínez, L., et al. Predicting gene expression from DNA sequence using deep learning models. Nat Rev Genet (2025). 

Please refer to the How to Apply section below before submitting your application.