Raphael Franzini, Ph.D.

DNA-encoded libraries: Streamlining drug discovery

One of the key steps in developing a new drug is to identify molecules that bind to a putative therapeutic target. However, there is a near-unlimited number of possible molecules and to identify the right one is a formidable challenge.

One approach to achieve this goal is to tag compounds with DNA strands whose sequence encodes for the structure of each compounds. In this way it is possible to use target proteins immobilized on surfaces as baits to fish for compounds that bind to this protein. Sequencing the attached DNA codes then allows to identify the corresponding molecules. In fact, this approach provides a semi-quantitative estimation of the target affinity of each compound in the library.

Such DNA-encoded libraries are nowadays used routinely in drug discovery at pharmaceutical companies. However, the prospect of this method remains largely untapped in academic medicinal chemistry efforts because of the costs associated with generating large library platforms and the validation of multiple hit compounds. We aim to overcome this problem by generating libraries that are structurally designed with specific protein families in mind. In this way, we can achieve consistent screening success at a fraction of the costs of large one-fit-all DNA-encoded library platforms.

Using such libraries, we were able to discover high potency hit compounds for several enzymes with speed and cost-efficiency unachievable by conventional methods. In parallel, we aim to further our understanding of how to synthesize and design such library and to develop algorithms for extracting important structural data from such library screens.