Machine Learning Models to Predict Chemotherapy Resistance in Breast Cancer Using Single-Cell Sequencing

Abstract

Chemotherapy is one of the main therapies for breast cancer, which is still one of the leading causes of cancer-related deaths globally. Chemotherapy resistance, which frequently results in treatment failure and a poor prognosis, is a major obstacle in the treatment of breast cancer. Early detection of chemotherapy resistance can greatly improve individualized treatment plans. In this work, we investigate how single-cell RNA sequencing (scRNA-seq) data might be used to predict chemotherapy resistance in breast cancer using machine learning (ML) models. Because cancer cells are heterogeneous, scRNA-seq offers a unique chance to identify genetic characteristics linked to treatment resistance at a fine level. Our goal is to use machine learning techniques to examine scRNA-seq data in order to find patterns and biomarkers that potentially indicate treatment resistance in breast cancer patients.

We preprocessed publically accessible scRNA-seq data to filter and normalise gene expression profiles, then employed dimensionality reduction and feature selection methods. We assessed the predictive power of a number of machine learning models, such as Random Forest (RF), Support Vector Machine (SVM), and Neural Networks (NN), for chemotherapy resistance. Accuracy, precision, recall, F1-score, and AUC-ROC were used to assess the model's performance. According to our findings, chemotherapy resistance may be reliably predicted by machine learning models; the Neural Network model had the highest AUC-ROC score. Furthermore, resistance was found to be significantly influenced by gene expression characteristics associated with immune response, cell cycle regulation, and drug metabolism. This work advances precision oncology by showing how single-cell sequencing and machine learning can be used to predict treatment resistance in breast cancer. The results imply that future clinical uses of ML models may play a significant role in customizing chemotherapy regimens for patients, enhancing results by preventing inefficient treatments.

Keywords:  , , , , , and .