Machine Learning for predicting yeast phenotype from genotype for biotech applications

This project will apply artificial intelligence approaches to address this challenge: given data on yeast genotypes, growth conditions and phenotypes (traits), can we develop predictive models for the phenotype of novel yeast strains and hence ultimately predict strains that could out-perform any of those in the training data.

Such novel strains could be produced using synthetic biology approaches and the model predictions tested.

Yeast is an ideal platform for the manufacture of biomedically important protein products, such as life-saving medicines.

The diversity of yeast genotypes and protein products means that the best strain for optimal yield of a given product is typically unknown - but ripe for identification using novel AI methods.