Huntington, C.; Vater, A.; Siegel, J.

Recent groundbreaking advances in protein structure prediction have significantly propelled the fields of protein design and engineering. However, improving computational enzyme modeling remains a key area of interest, with exciting opportunities to enhance predictive capabilities. Existing algorithms face a crucial hurdle: predicting enzyme stability and function with accuracy. To address this limitation, large sets of experimental data that capture enzyme structure-function are needed to guide the training and testing of next-gen tools. The Design 2 Data (D2D) program aims to build such a data set by engaging students from around the world to contribute standardized experimental data of enzyme variants. The flagship enzyme dataset of D2D, beta-glucosidase B (BglB), currently contains over 1300 records. The data consists of kinetic (kcat, KM, and kcat/KM), and thermal stability measures to expand the range of data available to train new computational algorithms. This paper explores six new single point mutants produced as part of this larger project and investigates variations in functional outcomes across the mutated sites. These data taken together with the larger and growing mutant library aim to deepen our understanding of the relationship between structure and function in BglB.