Related Experiment Video
Updated: Apr 18, 2026
Implementation of In Vitro Drug Resistance Assays: Maximizing the Potential for Uncovering Clinically Relevant Resistance Mechanisms
Published on: December 9, 2015
Integrating computational chemistry and machine learning to predict KRAS mutation-induced resistance
Katarzyna Mizgalska1,2, Konstancja Urbaniak3, Denis J Imbody4
1Department of Machine Learning, Moffitt Cancer Center, Tampa, FL, United States of America.
Abstract:
Mutation-induced drug resistance is a major contributor to the failure of targeted cancer therapies, particularly in tumors driven by mutations in the KRAS oncogene. Although covalent inhibitors effectively target KRAS G12C, secondary mutations such as G12C/Y96C, G12C/Y96S, and G12C/Y96D lead to resistance despite leaving the covalent attachment site intact. To predict these resistance outcomes, we developed a computational framework that integrates molecular dynamics-derived structural, energetic, thermodynamic, and contact-based descriptors with machine learning. Features extracted from simulations of treatment-sensitive and treatment-resistant KRAS mutants were used to train logistic regression, random forest, support vector machine, and Bayesian Network classifiers, achieving average accuracies above 90%. Solvent-accessible surface area variability, Lennard-Jones 1,4 energy, mean square displacement, and root mean square fluctuation emerged as the most discriminatory features. Residues G10, E62, and H95 showed the highest predictive value. This approach highlights conformational and solvent-exposure changes as central drivers of KRAS drug resistance and provides a generalizable workflow for other clinically relevant mutant targets.
Insights
Drug resistance in KRAS-mutant cancers can be predicted using a computational framework. This approach integrates molecular dynamics and machine learning to identify key protein changes driving resistance, achieving over 90% accuracy.
Area of Science:
- Oncology
- Computational Biology
- Biochemistry
Background:
- Mutation-induced drug resistance is a significant challenge in targeted cancer therapies, especially for KRAS-driven tumors.
- Secondary mutations in KRAS (e.g., G12C/Y96C) can confer resistance to covalent inhibitors by altering protein dynamics, even when the drug binding site remains intact.
Purpose of the Study:
- To develop a computational framework for predicting drug resistance in KRAS mutants.
- To identify structural and dynamic features that distinguish treatment-sensitive from treatment-resistant KRAS variants.
Main Methods:
- Integration of molecular dynamics simulations with machine learning algorithms (logistic regression, random forest, support vector machine, Bayesian Network).
- Extraction of molecular features including structural, energetic, thermodynamic, and contact-based descriptors from simulation data.
- Training and validation of predictive models using data from sensitive and resistant KRAS mutants.
Main Results:
- The computational framework achieved average accuracies exceeding 90% in distinguishing resistant from sensitive KRAS mutants.
- Key predictive features included solvent-accessible surface area variability, Lennard-Jones 1,4 energy, mean square displacement, and root mean square fluctuation.
- Residues G10, E62, and H95 were identified as critical sites influencing drug resistance.
Conclusions:
- Conformational and solvent-exposure changes are central mechanisms driving KRAS drug resistance.
- The developed computational workflow provides a generalizable strategy for predicting and understanding mutation-induced resistance in KRAS and other cancer targets.
