Research Goal

The underlying scientific goal of D2D is to facilitate “academic crowd-sourcing” to rapidly address protein design questions that would normally take isolated labs decades to answer.

Our challenge

To support a human population approaching eight billion, we need to be able to efficiently create never-before-seen proteins for novel therapeutics, and industrial and agricultural applications. Computational protein design allows researchers to explore possibilities at an unprecedented speed; however, a grand challenge that faces protein engineering communities is a lack of functional data for software benchmarking modeling algorithms.

A team-based solution

In 2016, the Siegel Lab published the largest enzyme functionality dataset of its kind with ~100 characterized mutants and the dataset was expanded upon in 2017. Based on this early scouting data, a critical result from the analysis of Carlin et al. (2017) indicated that more data are required before advanced computational methods can be used to start quantitatively predicting function. Given the vast potential sequence space of proteins, (e.g., for beta-glucosidase there are ~10,000 single-point mutations) the D2D project is ideal for the work of many hands and many minds to investigate the effects of structural mutations in an enzyme system on soluble protein expression, thermal stability, and catalytic efficiency. The quantitative data D2D students generate directly feeds back to the protein modeling research community for critical, predictive algorithm training!