Identification of Potential Breast Cancer Drug through an Integrative Approach of Single-Cell RNA Sequencing Data Analysis and Machine Learning

Abstract

This abstract presents a study that aimed to identify potential drugs for breast cancer treatment by analyzing large datasets and employing computational methods. Initially, breast cancer data from two datasets, Cancer Cell Encyclopedia (CCLE) and Genomics of Drug Sensitivity in Cancer (GDSC), were combined and tested for drug sensitivity using a drug-dose response curve. The drugs showing sensitivity in all eleven breast cancer cell lines were shortlisted for further evaluation. ScRNA-Seq analysis identified five drugs overlapping with the drug-sensitivity data, namely Afatinib, Bortezomib, Gemcitabine, Navitoclax, and Trametinib, which were selected for further investigation. Afatinib was evaluated based on its IC50 value and consistency across eleven cell lines. The target of Afatinib was identified from the datasets and subjected to machine learning using a Python script, which provided predictions of similar molecules in the ChEMBL database. Docking analysis was performed using the PDB ID 4G5J as a reference to predict the interactions of the target. The docking score for Afatinib was compared to a resultant compound, ChEMBL233325, with a high score of -9.4. The test compound ChEMBL233324 was selected for MD simulation as it consisted the same residue as our control compound. Overall, this study employed computational methods and analysis techniques to identify potential drugs for breast cancer treatment, providing valuable insights for further research and development.