Machine learning approaches to biosynthetic pathway optimisation

Lead Research Organisation: Imperial College London
Department Name: Life Sciences


We will test the hypothesis that machine learning can provide an effective and efficient method to find optimal solutions to complex biological design problems. We have developed a framework for the design and build of biological systems based our BASIC DNA assembly process. Automation of this has led to a step-change in our ability to rapidly prototype new biological designs: the vastly reduced timescale and cost afforded by this development therefore presents an opportunity to address new approaches to biological optimisation. The BASIC framework provides a conceptual design space for the construction of operons and their control through regulatory elements: promoter, 5'UTR, RBS, gene order, copy number. This design space also provides the framework for learning, since the DNA modules assembled also represent the parameter space for data analysis. This enables closed loop learning and iterative optimisation.

Here we will apply this automation and assembly framework to address the problem of biosynthetic pathway optimisation. Due to the intricacies of a biosynthetic pathway, the output will be contingent on a number of important, but unknown, factors. The optimum design to maximise yield of a biosynthetic product thus cannot be predicted, because the design rules for any particular biosynthetic pathway cannot be known in advance. Here we seek to develop a semi-autonomous approach that will use inductive logic programming to develop and test hypotheses by guiding the experimental plan.


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S022856/1 31/03/2019 29/09/2027
2294175 Studentship EP/S022856/1 30/09/2019 29/09/2023 Liam Hallett