PhD in Chemistry: Predicting the Kinetic Landscape of a Model Enzyme by Machine Learning

The optimisation of various valuable properties of biological catalysts such as selectivity and catalytic efficiency are possible by the application of “directed evolution” which has been recognised in the award of the Nobel prize for chemistry this year. However the correlation of kinetic activity, both the raw kinetic efficiency (Kcat) and the affinity for substrate (partially described by Km) to either structure or sequence is recognised as a difficult problem. We have shown that the microbial type II dehydroquinase family of enzymes (a well-established target for antibiotic development) is an excellent model system:
• Differences in catalytic activity spanning 4 orders of magnitude.
• Trivial and reliable protein expression and purification
• Simple 1 substrate kinetic UV assay
• The short 150 amino acid sequence and the availability of >50 diverse clones in house. Reference sequence data on >20,000 unique protein sequences is available.
• The proteins are amenable to crystallisation and we have 20+ crystal structures to atomic resolution obtained.

The project commences in October 2019.

The aim of the project is to correlate the kinetic, sequence and structural data in such a way that whole protein family can be understood and the potential kinetic landscape predicted. Predictions from any machine learning algorithm based on the available data can be tested trivially by experimentation. The aim is to develop methodology to compare and contrast methods using amino acid sequences and 3D structures/active sites to allow prediction of catalytic activity.

Overview of Research
The PhD research project will involve some routine lab based protein expression, purification, kinetic characterization and crystallization. Site directed mutagenesis and chimeras will be produced using simple kit based molecular biology techniques. In addition the project will involve significant computational and bioinformatics components. The student will use existing programs to determine, refine and analyze crystal structures. In addition the use and development of tools and methods to convert 3D data into suitable matrix based 2D data and correlate these with both sequence and kinetic data will be key to the success of this project.

The ideal candidate will hold (or expect to be awarded) a 1st Class or Upper 2nd Class degree in a suitable science subject. Ideally the student should have programing and computing experience. Enthusiasm and good problem-solving abilities are essential.

How to Apply: Please refer to the following website for details on how to apply:
http://www.gla.ac.uk/research/opportunities/howtoapplyforaresearchdegree/

Prospective candidates for the studentship should contact Dr Lapthorn directly by E-mail ([Email Address Removed]) as soon as possible and supply a CV and covering letter. Applications will be considered on a rolling basis.