Julie Wilson

Profile

Biography

Julie Wilson is a professor in Applied Statistics in the Department of Mathematics at the University of York. Her research interests lie in the application of mathematical modelling and statistical methods, often applied to large biological and chemical data. She has extensive experience in chemometric method development and statistical pattern recognition, including multivariate analysis, classification and machine learning techniques. Current industrial collaborations involve Fera Science, AstraZeneca, GlaxoSmithKline, Croda and Eluceda.

Departmental roles

• Impact Officer

Research

Overview

My research interests lie in the application of mathematical modelling and statistical pattern recognition methods to biological and chemical problems. Industrial collaborators include large pharmaceutical companies, AstraZeneca and Glaxo Smith Kline, as well as small local businesses with projects ranging from quality control of hair dyes to optimisation of maggot diets, with insects being farmed as a sustainable source of protein for animal feed (pigs, chickens and fish).

Collaboration with the Food and Environment Agency, now Fera Science Ltd, has led to projects involving food safety, plant and animal health and environmental issues. We have developed novel chemometric and bioinformatic methods in response to issues arising in various analyses, including the study of drought and disease resistance in plants, the authentication of Manuka honey and the search for biomarkers for Mad Cow disease. An example is the method developed to deal with differences introduced by acquiring liquid chromatography-mass spectrometry (LC-MS) data in batches to allow necessary calibrations and cleaning of the instrument. We developed a correction method that identifies the trend over time and can highlight differences between experimental groups previously hidden by instrumental variation. The code to perform batch correction can be found here.

In contrast to the mega-variate data sets obtained by -omics technologies, the analysis of images often involves the extraction of relatively few relevant features from huge numbers of examples. Imaging projects include the analysis of cell morphology for discrimination of cell types, identification of bladder cancer cells and investigation of cells’ response to drug treatment. 

Protein crystallisation is of fundamental importance in structural biology and I have been involved in international efforts to improve crystallisation strategies through automated analysis of images from crystallisation trials.

The pH of an experiment is an important parameter and various buffers are used to maintain a specific pH, but the pH of the buffer solution and can be highly inaccurate. Using data from thousands of experiments provided by the Bio21 Collaborative Crystallisation Centre (C3) in Australia, we have been able to model the combined effect of other chemicals in the crystallization solution to give a better estimate of the true. The pH for the chemical ‘cocktail’ can be predicted here.

Research group(s)

Mathematical Biology and Chemistry Research Group

Supervision

PhD Supervision

PhD research projects I supervise tend to be interdisciplinary and involve developing new methods to extract useful information from biological or chemical data.

Current Research Students

• Jamie Milne - jm2163@york.ac.uk
• Laura Wiggins, co-supervised in Biology, “Using machine learning to predict cell type and fate in heterogeneous populations.”

Past Research Students

• Richard Davis, “Analysis of pattern recognition techniques applied to 1H NMR metabolomics data.”
• James McKenzie, co-supervised in Chemistry, “Assessment of the complementarity of data from multiple analytical techniques.”
• Jobie Kirkwood, co-supervised in Computer Science, “Analysis of protein crystallization parameters.″
• Joanna Simpson, co-supervised in Chemistry and Archaeology, “Investigating the relationship between glutamine deamidation and collagen breakdown in bone.”
• Martin Ruscilowicz, co-supervised in Computer Science, “Computational tools for the processing and analysis of time-course metabolomic data.”
• Christopher Saunders, co-supervised in Biology, “The role of the chemokine receptor CCR5 during the infection of macrophages by Leishmania parasites.”
• Guy Beavis, “Chemometric method development for metabolomic analyses for food security and authentication.”
• George Clarke, co-supervised in Chemistry, ″Intelligent synthesis: a unique reaction optimisation strategy.”

Teaching

Undergraduate

• Statistical Pattern Recognition
• Practical Data Science with R