Defense Date
2026
Document Type
Dissertation
Degree Name
Doctor of Philosophy
First Advisor
Dayanjan S Wijesinghe
Abstract
More than ninety percent of multi-kinase inhibitor programs fail in clinical development, most often from off-target toxicity, yet the discontinued compounds, and the rich pharmacological data they carry, are rarely reused. This dissertation presents an integrated, open computational platform that mines failed multi-kinase inhibitors and reframes their off-target binding profiles from liabilities into experimentally testable repurposing hypotheses.
Using CDK7 (ChEMBL CHEMBL3055) as a case study, the platform demonstrated that clinical failure tracks engagement of cancer-essential kinases rather than raw promiscuity, translated this distinction into a Repurposing Readiness Score and an A/B/C triage, and applied a fragment-based genetic algorithm to generate novel chemotypes with a predicted, ordinal reduction in CDK7 binding, supported concordantly by XGBoost IC50 prediction and molecular docking. Preliminary transferability was demonstrated for GSK3B in Alzheimer's disease.
All results are computational predictions requiring experimental validation, the platform's contribution is a transparent decision-support framework that prioritizes discontinued kinase inhibitors as testable repurposing hypotheses rather than finished therapeutics.
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
7-21-2026