The overall goal of my CPRIT proposal is to leverage my training and expertise in computational sciences and structure-based molecular modeling to mine the vast cancer data resources and billions of compounds in DNA-encoded chemical libraries (DEL) to discover new cancer therapeutics. I will combine deep learning techniques and gene expression-based DEL design to explore the enormous chemical space that is not covered by existing DEL. Initially, I will focus on small molecule specific compounds to two protein kinases, the mutant epidermal growth factor tyrosine kinase receptor with exon 20 insertions (EGRF-Ex20ins) and mitogen-activated protein kinase 5 (MAP2K5, a dual specificity kinase als...

Read More

The overall goal of my CPRIT proposal is to leverage my training and expertise in computational sciences and structure-based molecular modeling to mine the vast cancer data resources and billions of compounds in DNA-encoded chemical libraries (DEL) to discover new cancer therapeutics. I will combine deep learning techniques and gene expression-based DEL design to explore the enormous chemical space that is not covered by existing DEL. Initially, I will focus on small molecule specific compounds to two protein kinases, the mutant epidermal growth factor tyrosine kinase receptor with exon 20 insertions (EGRF-Ex20ins) and mitogen-activated protein kinase 5 (MAP2K5, a dual specificity kinase also known as MEK5). My parallel strategy will accelerate the discoveries of hit compounds that can overcome resistance to tyrosine kinase inhibitors (TKIs) and place my group at the forefront of computational cancer drug screening. The Aim 1 objective is to use in-house DEL (>6 billion compounds) to identify hits that bind and inhibit our target proteins. The Aim 2 objective is to use the entire DEL selection output, statistical binding scores of billions of compounds and their structural information, to predict the target binding propensity of building block combinations and of commercially available compounds. The Aim 3 objective is to design and synthesize target focused DEL by incorporating building blocks derived from analyzing drug-induced gene expression alterations. By matching the drug-induced gene expression profiles to genetic knock-down induced gene expression profiles, building blocks with potential to mimic the knock-down effect of EGFR-Ex20ins and MEK5 will be used to build focused chemical libraries. Combining deep learning techniques and gene expression-based chemical library design will be the foundation of my next-generation DEL platform. This novel technology will accelerate the discoveries of new drugs for targeted cancer therapeutics.

Read Less