Using cell free systems and machine learning to prototype engineered protein designs

点击翻译按钮获取中文摘要

Cell-free systems have recently been developed as a prototyping platform for synthetic biology designs. The ability to screen a large number of genetic designs prior to transformation into a living chassis speeds up the design cycle and provides initial validation for predicted functionality. In this project, a collaboration with the UK synthetic biology company LabGenius will aim to develop a machine learning platform for protein engineering applied to a number of target proteins including biotherapeutics. The overall aim is to establish a cell-free workflow to allow the rapid assessment of engineered protein designs in vitro. A major challenge will be to ensure that the target proteins can be produced at a quality, yield, price point and purity required for down-stream testing. Another major aim of the project is to generate large datasets on specific target proteins which can then be used to inform and test the application of machine learning (ML) algorithms. A major challenge for this aim will be to ensure that the experimental platform is sufficiently robust to produce consistently reproducible results. If the datasets produced are too noisy, it will be very difficult to effectively use ML. We plan to use a number of non-commercially sensitive target proteins including an off-patent single chain antibody and industrially relevant enzyme (with available 3D structural information) to explore the validity of applying cell free measurements to assess engineered protein designs that are derived by machine learning algorithms. The technologies required for the core components of the proposed platform are already in place. The Freemont lab has already developed a low volume 384 well plate cell free assay platform that utilises the Labcyte acoustic liquid handling robot which will provide the basis of the proposed project. The cell-free prototyping platform is located in the London Biofoundry which Prof Freemont is co-director (https://www.londonbiofoundry.org/). Dr Baldwin has developed platforms for automated DNA assembly that will be used to construct the libraries of genetic variants. By assessing the specific known genetic constructs assembled, ML approaches developed by LabGenius will be used to iterate the design cycles.