Machine learning-generated compression modulus database for 3D printing of gelatin methacryloyl

3D bioprinting enables the fabrication of printable tissues, including those for neural, cartilage, and skin repair. The mechanical properties, especially stiffness, of 3D-bioprinted scaffolds are crucial for modulating cell adhesion, growth, migration, and differentiation. The stiffness of a scaffold can be adjusted post-printing by modifying the hydrogel concentration, crosslinker concentration, light intensity during photocrosslinking, and duration of crosslinking. The optimization of these conditions to produce the desired scaffold stiffness for a particular cell type or application is a time-consuming and rigorous process. This study developed an innovative approach to predict the compression modulus of 3D-printed gelatin methacryloyl (GelMA) scaffolds using the Bayesian optimization (BO) algorithm. Through just 10 iterations (75 experimental data points), the model was able to predict > 13,000 possible compression modulus values in a search space comprising four independent variables (GelMA concentration, crosslinker concentration, ultraviolet light UV distance, and UV exposure time). This approach can be utilized in other photocrosslinkable bioinks for 3D printing that have a myriad of pre- or post-printing parameters that can affect scaffold stiffness.