Cheminformatics Analysis of DNA-encoded Libraries

Description:
Rapid exploration of chemical space can be achieved through exponentially growing DNA-encoded libraries (DELs). In DELs, library members are assembled by piecing together small molecules, known as building blocks (BBs), in different combinations. The synthetic history of each library member is then stored in unique DNA barcodes, allowing millions of compounds to be screened and identified in single experiments.

Predicting active compounds from existing DEL data is a problem well-suited for computational models. We can rapidly generate large libraries and identify active compounds to feed as training data. Instead of requiring models to predict the behavior of compounds they have never seen before, we can search for novel compounds by proposing new BBs to add to the DEL or by combining existing BBs in new ways.

In this talk, we discuss the data challenges of processing experimental DEL data and insights that can be used to guide future computational analyses of DELs. We introduce a procedure that combines physical methods with clustering techniques to analyze DELs at the building block level. Lastly, we provide an outlook on how advances in DEL automation can be combined with computational models to accelerate drug discovery in the future.

Speaker: Chris Zhang - University of California, Irvine
Chris is a 5th year graduate student working in the lab of Professor David Mobley at UC Irvine. His research focuses on developing computational methods to advance pharmaceutical drug discovery. He also has an interest in data science and has taken on a variety of cheminformatics projects during his graduate career.