Hosted by the Bioinformatics Research Group

Abstract: Integral membrane proteins constitute ~25% of all cell proteins and are critically important for diverse processes including solute and ion transport, intracellular signalling, cell communication, and bioenergetics. However, both the fundamental understanding and commercial exploitation of membrane proteins remain plagued by difficulties in their recombinant production. The unpredictable (or entirely failed) recombinant expression of membrane proteins impacts basic research including structural biology, enzymology, protein engineering, biodesign, and cell biology; in applied research this same expression problem stifles innovation across drug discovery, biologics, biotechnology, synthetic biology, biomanufacturing, and bioprocessing. We now propose to address this research challenge by establishing a predictive computational model that interprets the molecular code underpinning membrane protein expression and defines the nature of the expression fitness landscape. I will describe our encouraging first steps towards such a model through a novel approach that integrates protein design with machine learning.

Light refreshments will be provided afterwards.