Quantitative Structure-Mutation-Activity Relationship Tests (QSMART) model for protein kinase inhibitor response...

Area of Science:

• Oncology
• Pharmacogenomics
• Computational Biology

Background:

• Protein kinases are crucial drug targets in cancer, but understanding their complex interactions with drugs is challenging.
• Personalized medicine relies on predicting patient response to kinase inhibitors, yet genomic and proteomic data integration remains a bottleneck.

Purpose of the Study:

• To develop an explainable computational model for predicting protein kinase inhibitor response.
• To identify key molecular features and interactions that determine drug sensitivity in cancer cell lines.

Main Methods:

• Proposed a multi-component Quantitative Structure-Mutation-Activity Relationship Tests (QSMART) model with neural networks.
• Integrated multi-omics data, including genomic mutations and drug structures.
• Utilized non-small cell lung cancer as a case study for EGFR inhibitors afatinib and lapatinib.

Main Results:

• The QSMART model accurately predicted drug responses in cancer cell lines.
• Identified significant interaction terms involving drugs, pathways (e.g., JNK apoptotic pathway), and mutant kinases.
• Highlighted the contribution of drug substructures and mutant residue properties to drug efficacy.

Conclusions:

• The QSMART model effectively integrates multi-omics data for accurate drug response prediction.
• The framework provides explainable insights into genotype-drug sensitivity relationships.
• The approach is extendable beyond kinase inhibitors to other cancer targets.