Alec McKinlay - The role of the human gut microbiome in cancer aetiology

After completing an undergraduate degree in Biochemistry and an MSc in Epidemiology, I was keen to continue my postgraduate studies at the University of Bristol. As a result of this, I applied for an ICEP PhD project, which aims to investigate the role of the human gut microbiome in cancer aetiology and its potential use as a predictive biomarker in cancer diagnosis. The microbiome, home to a range of bacteria and other organisms, varies between individuals due to pre-determined characteristics, i.e. our genetics, and modifiable factors including diet, lifestyle, and pro/prebiotic consumption. These differences are shown to mediate a range of pathologies, including cancer. The project combines both lab-based biochemistry elements along with an epidemiology computational approach to triangulate findings and help disentangle links between the microbiome and cancer.

 

Amy Francis - Functional effects of genetic variants

My project uses machine learning methodologies to predict the functional impact of genetic variants in clinical genome sequencing. In particular, our work builds on previously developed tools such as CScape and FATHMM. Algorithm development will be attained through modern machine learning libraries such as Scikit-Learn, Tensorflow, PyTorch, in conjugation with the python programming language. We also plan to combine these tools with modern approaches to drug repurposing. Such a tool could be highly beneficial in a clinical setting and could facilitate precision medicine.

 

Becky Scanlan - The role of the human gut microbiome in cancer aetiology

It was during my Biology and Nanomedicine studies involving projects developing both insect models and in vitro co-culture responses to carcinogenic immune insults that led to my increasing interesting in systemic cancer risk factors. This led me to pursue an ICEP-aligned PhD with Cancer Research UK at the University of Bristol, drawn to the focus on large-scale population studies and implementation of impressively robust causal inference methods within the department. By utilising genetic variants and cutting-edge statistical techniques in an epidemiological framework, I am investigating associations between human gut microbiome traits and breast cancer incidence and progression. My project will involve an array of triangulation techniques including observational, one- and two-sample methods, and appropriate sensitivity analysis plus corroboration with in vivo murine studies.

 

Emma Hazelwood - The impact of the metabolic environment of obesity on cancer promotion

My research focuses on the colorectal-specific effects of obesity, and how these could then increase the risk of colorectal cancer. Specifically, my research combines genetic epidemiological and cell biology approaches to look at how gene expression changes caused by obesity alter colorectal cellular metabolism. Some of these changes may mediate the effect of obesity on colorectal cancer risk, and represent potential targets in disease prevention.

  

George Richenberg - Identifying new targets for cancer prevention and therapy by linking common exposures or cancer risk factors to tumour molecular mechanisms

Cancer risk factors, such as smoking and ultraviolet light, are well known drivers of malignancy but the impacts of everyday endogenous and exogeneous factors, such as adiposity, diet and insulin resistance, on cancer onset and progression are poorly understood. During this PhD, computational and epidemiological approaches will be used to analyse multi-omic data from The Cancer Genome Atlas (TCGA) to causally associate common exposures with tumour molecular features. A combination of cutting-edge statistical approaches and multivariable regression modelling techniques, including expression quantitative trait locus analysis, Mendelian randomisation and polygenic risk scoring methods, will be adopted across a pan-cancer model of hormonal cancers in breast, ovarian and prostate tissues to determine how these exposures associate with cancer-driving mechanisms across cancer types. Causally linking risk factors with carcinogenic molecular mechanisms will aim to provide specific molecular targets to facilitate advancements in precision-based cancer treatments to help reduce cancer’s morbidity and mortality. Identified causal risk factors may also help to elucidate biomarkers for early cancer diagnosis and early therapeutic intervention to improve patient survival. 

 

Luke Mahoney - Using genetics to separate adiposity from hormonal changes at menopause and investigate their causal relationship with cancer.

My PhD project is using genetics and mendelian randomisation to understand the role played by hormonal and adiposity changes that occur during menopause. The project will aim to identify and characterise genetic variants associated with age at natural menopause, sex hormones and fat distribution. I will then look at the causal effect of hormonal and adiposity changes and specifically their interaction. I will use multiple hormone measures and definitions of adiposity to capture a full picture of the causal effect of these changes.

 

Olympia Dimopoulou - Drug repurposing for prostate cancer prevention

I am a biologist with a MSc in Biostatistics trained at the National and Kapodistrian University of Athens. In 2020 I joined the ICEP team as a Research Associate in Aetiological Epidemiology. I am currently working on my PhD project exploring mechanistic pathways associated with prostate cancer (PCa) in a triangulation framework and detecting existing licensed drugs and drug targets that could possibly reduce PCa risk and progression. The reasoning behind this project is that licenced drugs could be repurposed to reduce PCa risk and development due to the interplay between their molecular targets and many biological pathways.

 

Scott Waterfield - DNA methylation indices of protein alterations, and early cancer detection

Throughout my project I will be creating models using around half a million to a million DNA methylation data points to predict the expression of numerous different proteins. Different types of machine learning models, such as penalized regression, random forests, and neural networks will be compared for their efficacy in predicting protein expression using DNA methylation data. Following this, these proteins prediction models will be integrated into cohort studies which include individuals who had blood taken in advance (ideally months/years) before a cancer diagnosis, which will allow us to determine if they can identify cancer from blood samples in individuals before they have been diagnosed, offering future hope for earlier diagnostic tests.