Non-conventional yeast strain optimisation for industrial protein production using deep learning

Synthetic biology has a potential to enable cheaper, cleaner, more economically and environmentally sustainable biomolecule production. Animal-free rennet (for cheese production), insulin (for medical purposes), and leghemoglobin (for alternative meat production) are all produced using a process called precision fermentation -- where a microorganism is modified into a mini-factory of a desired molecule. However, despite massive advances in precision fermentation, it remains difficult to achieve sufficient production of some proteins, resulting in high production costs and limiting their beneficial impact for businesses.

Precision fermentation has a number of techniques to improve production, but these techniques are labour-intensive and limited by our current understanding of regulatory sequence coding principles. In this project, we will evaluate the potential of _in silico_ optimization for protein production. We will validate machine-guided sequence optimization for the industrial production of a high-value protein. Like many recombinant proteins, its production costs are driven by a complicated production process in prokaryotes and costly purification. Yeast production systems have been shown to be more advantageous for such protein production but their titers are not yet on par with _E. coli_. The purpose of this project is to improve the cost efficiency of these yeasts for high-value protein production.

More broadly, our project will provide a pathway to the development of economically and environmentally sustainable production chassis for protein ingredients, critical to food, cosmetics, and pharmaceuticals markets. By using precision fermentation to create these ingredients, we can reduce the carbon footprint of the production process, as well as address issues of food security and public health. Moreover, we hope that our pipeline will open up novel avenues for precision fermentation of other high-value proteins that currently have no means of scaled manufacturing.

Our proposal seeks to overcome the limitations of current practices and leverage evolving technologies to accelerate the development of new expression systems for biomanufacturing. We aim to create an _in silico_ optimization platform that can drive innovation in this field and solve major production bottlenecks, significantly contributing to the growth of the biotechnology sector in the UK.