PhD Position in “Novel biomarkers for human gut AMPK activation using AI-driven multi-omics analysis”

Overview

This project aims to discover and validate a panel of blood-based biomarkers in humans for measuring activation of AMPK, a protein complex which acts as a detector of cellular “fuel deficiency”. The core strategy involves a systematic, AI-driven analysis of existing and new multi-omics data to pinpoint a precise molecular signature.

Supervisors: Professor Gary Williamson (School of Biological Sciences) and Dr Shuyan Li (EEECS).

Timeliness and importance:

Obesity, overweight, and type 2 diabetes are rapidly growing worldwide. There are various strategies but initial weight loss is hard to maintain, even after drug treatment. A potential target is AMP-activated kinase (AMPK), a protein complex which acts as a detector of cellular “fuel deficiency”. Importantly, clinical AMPK activators act as calorie restriction mimetics and lower the risk of chronic conditions. During low energy, AMPK phosphorylates specific enzymes and growth control nodes to increase ATP generation and decrease ATP consumption, and the activation capacity declines in metabolic stress and with aging. Metformin is a clinical AMPK activator, and the activation may be potentiated by the drug salicylate. However, there are many benefits in a dietary strategy, but research is hindered by a lack of direct biomarkers of AMPK activation for use in human dietary intervention studies.

Research question:

Generic energy metabolic status can be estimated indirectly using blood metabolomics and other ‘omic techniques on subjects with obesity, diabetes and after exercise. The research question is: can we develop a biomarker which could more specifically measure AMPK activation in humans using blood samples.

Objectives:

1. Aggregating metabolomic, proteomic, phosphoproteomic, and transcriptomic datasets from previous human, animal, and cell studies involving known AMPK activators (e.g., metformin, salicylate, dietary phytochemicals) and controls. These diverse datasets will be integrated into a centralized database where raw and processed data are stored in structured formats, meticulously annotated with experimental metadata.

2. The AI-driven analysis will then proceed in three key stages.
(a) supervised machine learning models, such as Random Forest and LASSO regression, will be employed to analyse this multi-omics input. Their goal is to perform classification and identify a shortlist of output features—changes in specific metabolites, protein phosphosites, and genes—that are most strongly predictive of AMPK activation.
(b) The robustness of these AI-identified outputs will be validated using unsupervised learning techniques like UMAP to confirm they can consistently cluster samples by AMPK status.
(c) These top candidate biomarkers will be integrated into a simple, interpretable model (e.g., a logistic regression formula) to generate a single, quantitative “AMPK Activation Score.”

3. This validated methodology and the resulting biomarker panel or score will be directly applied to existing or new samples from human intervention studies. This will confirm the utility of these AI-identified outputs for assessing the effects of nutritional interventions, foods, and other compounds on AMPK activation and energy utilization in human subjects.

Data availability and collection: Initial data to be analysed by AI is already published and available. Human blood samples will be developed in a parallel project funded by the supporting company.

Academic Requirements:

The minimum academic requirement for admission is normally an Upper Second Class Honours degree from a UK or ROI Higher Education provider in a relevant discipline, or an equivalent qualification acceptable to the University.

Desirable academic requirements:

A strong background in nutrition, biochemistry/biological sciences, and/or extensive experience of Python, background information of machine learning.

Funding Information

This project is part of the NILab Programme. Further details can be found here: https://www.qub.ac.uk/sites/nilab/

Applications are welcome from Home and International applicants. Please note that only a small number of NILab awards are available for international applicants. These international awards will be competitively allocated across all NILab projects based on the overall strength of the applications.