17 April 2023 in ChemRxiv
DNA-encoded libraries (DELs) provide the means to make and screen millions of diverse compounds against a target of interest in a single experiment. However, despite producing large volumes of binding data at a relatively low cost, the DEL selection process is susceptible to noise, necessitating computational follow-up to increase signal-to-noise ratios. In this work, we present a set of informatics tools to analyze DEL selection data so that subsequent DEL screens probe productive regions of chemical space. Our approach segments DEL data at the individual building block level to identify productive building blocks in a library. We show how similar building blocks have a similar probability of binding, which we then employ to predict the behavior of untested building blocks. Lastly, we build a model from the inference that the combined behavior of individual building blocks is predictive of the activity of an overall compound. We report a performance of more than an order of magnitude greater than random guessing on a holdout set, demonstrating that our model can serve as a baseline for comparison against other machine learning models on DEL data.