For this year's program, we used a total of 112 parts, including 48 basic and 64 composite parts. Most of these components were designed by ourself, and all of them have been tested to insure their reliability. You can click here to learn more about how we design, optimize these parts, or click on the part numbers when browse website to learn more about our programs.

Our favorite composite parts are BBa_K5157103, BBa_K5157098, and BBa_K5157114, for they can most significantly improve the degradation efficiency of PETase.

We used both LSTM and graph neural network (GCN) models to optimize the prediction of high-affinity PET-binding peptides. For further improvement, we used Mutcompute-super to introduce beneficial mutations at specific sites, in order to improve the expression level of the fusion protein and the degradation ability of the microplastic substrate by molecular modification.

After performed PET-binding peptide prediction with the help of deep learning, we then fused PETase with the efficient binding peptide obtained from the prediction and screening through linker to construct a fusion protein, so as to improve the binding ability of PETase to PET microplastics with the help of PET-binding peptide, thus improving the degradation efficiency. We also characterized the binding ability of PET-binding peptides to PET plastics with the help of eGFP.