Exploring Chemical Space for the Discovery of New Functional Materials and Therapeutics

Given the combined rate of advancement in both computational power and target identification for the treatment of high impact disease, computer-aided drug design (CADD) has emerged as a new alternative to traditional small molecule design, along with the advances in novel materials and biologic derived therapies. Paired with renowned contemporary methodologies for structural determination, the modern drug discovery toolkit has greatly expanded with both large industry and the average patient serving to benefit. Despite these advances, limits and biases in both data diversity and methodology execution still restrict the potential of many computationally guided processes. Designing new small molecules and biologics for previously untreatable targets is thus essential for continuing this path of discovery. The first portion of this dissertation explores the synthesis of sequence defined dendrimers, a new class of functional materials. Sequence defined dendrimers benefit from many of the known desirable properties of traditional dendritic systems, combined with a programable motif to explore sequence specificity. Sequence defined dendrimers with up to six peripheral sites have been successfully synthesized, with applications to combinatorial libraries via automated SPPS ongoing. The latter portion of this dissertation concerns the development of ChemHopper, a computational package which leverages novel molecular design for generative lead optimization. ChemHopper utilizes multi-scale R-group enumeration for the creation of an ultra-large hierarchical library of drug-like fragments. This library is procedurally generated to efficiently explore the chemical space of drug-like moieties, thus avoiding the bias included in open-source data derived from limited known hits. Chemhopper utilizes CPU/GPU parallelization to then screen its internal library for the production of new drug-like molecules, at speeds rivaling modern HTVS methods.