Projects with Fujifilm Diosynth Biotechnologies

This year, our CDT is collaborating with Fujifilm Diosynth Biotechnologies as part of their Centre of Excellence in Bioprocessing.

We have two extra projects on offer this year, developed and delivered in collaboration with Fujifilm Diosynth Biotechnologies.

Please note the projects are only available to applicants with UK home fee status.

If you're interested in these projects, you can apply just for them or for both the projects and the general CDT pool.

Please review the information below and reach out to the supervisory teams to discuss the projects before applying.

 

Supervisory team: Karl Burgess Karl.Burgess@ed.ac.uk, Diego Oyarzun D.Oyarzun@ed.ac.uk 

Hypothesis: Can harmonising analysis and biochemical interpretation in metabolomics allow better understanding of the underlying biology?

Project Description: Metabolomics is the ‘omics discipline closest to revealing the phenotype of an organism or biological system, and is used in disciplines as diverse as clinical analysis and biotechnology. Metabolism itself, however, is composed of hundreds of interlocking transient reactions mediated by enzymes with varying physicochemical properties and kinetics. Metabolites themselves have a huge variety of intrinsic properties and constraints on their ability to be analysed. This leads to significant issues with performing and interpreting metabolomics analysis: first, the key metabolites may be challenging to detect using the methodology selected, and second, the biological meaning or interpretation of the data is difficult to confidently assess, especially for typical ‘metabolite pool’ measurements. Put simply, if we detect that the abundance of a metabolite has increased, what does that mean, biologically? Does it mean that the pathway that metabolite belongs to is being increasingly utilised, thus the organism is making more? Or that it’s not used at all, so the pool is static.

This PhD proposal aims to explore these themes using machine learning and graph analysis to: predict ionisation in mass spectrometry and retention on chromatography of metabolites based on their physicochemical properties; identify focal points in metabolic networks where detection of a metabolite provides greater information, e.g. metabolites that are involved in pathway interactions; and finally to explore the concept of metabolic flux and how it relates to static metabolic measurements. We have previously demonstrated retention time prediction for metabolites in hydrophilic interaction liquid chromatography (HILIC), as well as the use of biochemical networks to allow probabilistic analysis of metabolism. This work extends and encapsulates these concepts to produce a holistic analysis of the meaning of the metabolome and our capability to measure it. 

The student will develop a combination of machine learning and practical mass spectrometry and chromatography skills. Specifically the maintenance and operation of high resolution accurate mass spectrometry, HILIC, reversed phase liquid chromatography, convolutional neural networks, graph networks and advanced statistical analysis. 

Proposed project outcomes:

1 – A tool that, given a set of seed compounds, allows prediction of not only column retention but also peak shape quality for chromatography.

2 – A tool that allows selection of a biochemical network, for example a genome-scale metabolic network for hepatocytes, Chinese hamster ovary cells or E.coli, and generates a list of accessible metabolites tailored to an analytical system that allow comprehensive, quantifiable analysis of the metabolic state of that network.

3 – A framework for analysis of metabolic data that allows biologists to confidently interpret metabolomics data in the background of an experimental design.


Supervisory team: Elise Cachat Elise.Cachat@ed.ac.uk, Filippo Menolascina Filippo.menolascina@ed.ac.uk

Project description pending


This article was published on