Despite years of effort and advances in technology, approaches able to bring a drug from bench to bedside remain costly, complex, time-consuming and without guarantee of success. The emergence of the computer-aided drug design (CADD) has improved every stage of drug development by expediting the process to 1-2 years and reducing the cost to $1.3 billion by generating more accurate and safer lead compounds. Still, there is an urgent need for the development of more powerful methodologies. Driven by this, multiple computational drug design methods were developed by us during the past few years, including 1) the target-specific molecular docking/virtual screening methods for identifying lead compounds, i.e., metalloproteinases-specific docking program MpSDock (Bai et. al., JCIM, 2015, 55:833-847), and DNA-specific docking tool iDNABinder (Bai et.al., NAR, 2017, 46:546-557); 2) the binding kinetic/thermodynamic prediction method, which is a generally applicable approach that enables the binding affinity and the binding kinetics to be accurately estimated (Bai et.al, PNAS, 2013,110:4273-4278), molecular binding mechanism to be revealed and a detailed strategy of binding-kinetics driven drug design to be proposed; 3) a druggable protein-protein interface identification method, named Fd-DCA, was developed by combing a fragment-docking method with evolutionary signals from sequence data (Bai et. al., PNAS, 113: E8051-E8058), which can accurately predict and characterize not only the druggable binding sites of PPIs, but also individual proteins, as well as provide fragment-based drug design strategy to create de novo small molecule or peptide drugs, or help to optimize lead compounds/peptides. By utilizing these methods, drug design applications were performed by targeting types of diseases related genes, e.g., breast cancer targets EGFR and NAF-1, Alzheimer’s disease target AChE, cancer target AP-1, and diabetic target DDPIV, etc. Several potent small molecular inhibitors and peptides were discovered and showed promising results. Moreover, couples of protein-protein interactions were identified and their binding complexes were constructed, i.e., the NAF-1/IASPP, CISD1/VDAC/CISD3, and CTP1A/Rab14 etc. These validation studies once again demonstrate the potential values of these computational methods in drug discovery and design area.

Improving the rationale for drug design by developing computational approaches

2019 Spring National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.