Related Experiment Video
Updated: May 16, 2025
Predictive Immune Modeling of Solid Tumors
Published on: February 25, 2020
Uncovering gene and cellular signatures of immune checkpoint response via machine learning and single-cell RNA-seq
Asaf Pinhasi1, Keren Yizhak2,3
1Department of Cell Biology and Cancer Science, The Ruth and Bruce Rappaport Faculty of Medicine, Technion - Israel Institute of Technology, Haifa, Israel.
Abstract:
Immune checkpoint inhibitors have transformed cancer therapy. However, only a fraction of patients benefit from these treatments. The variability in patient responses remains a significant challenge due to the intricate nature of the tumor microenvironment. Here, we harness single-cell RNA-sequencing data and employ machine learning to predict patient responses while preserving interpretability and single-cell resolution. Using a dataset of melanoma-infiltrated immune cells, we applied XGBoost, achieving an initial AUC score of 0.84, which improved to 0.89 following Boruta feature selection. This analysis revealed an 11-gene signature predictive across various cancer types. SHAP value analysis of these genes uncovered diverse gene-pair interactions with non-linear and context-dependent effects. Finally, we developed a reinforcement learning model to identify the most informative single cells for predictivity. This approach highlights the power of advanced computational methods to deepen our understanding of cancer immunity and enhance the prediction of treatment outcomes.
Insights
Machine learning predicts cancer treatment response using single-cell immune data. An 11-gene signature and advanced models improve prediction accuracy for immune checkpoint inhibitors.
Area of Science:
- Immunology
- Computational Biology
- Oncology
Background:
- Immune checkpoint inhibitors (ICIs) have revolutionized cancer therapy but show variable patient responses.
- Predicting patient response to ICIs is challenging due to tumor microenvironment complexity.
Purpose of the Study:
- To develop a machine learning model for predicting patient response to ICIs using single-cell RNA-sequencing data.
- To identify a predictive gene signature and understand gene interactions within the tumor microenvironment.
Main Methods:
- Applied XGBoost and Boruta feature selection to single-cell RNA-sequencing data from melanoma patients.
- Utilized SHAP values for interpretability of gene-pair interactions.
- Developed a reinforcement learning model to identify informative single cells for prediction.
Main Results:
- Achieved an AUC score of 0.89 for response prediction after feature selection.
- Identified an 11-gene signature predictive of response across multiple cancer types.
- Uncovered complex, context-dependent gene-pair interactions influencing predictivity.
Conclusions:
- Advanced computational methods, including machine learning, can significantly enhance the prediction of ICI treatment outcomes.
- Single-cell resolution and interpretability are crucial for understanding cancer immunity and treatment response.
- The identified gene signature and predictive models offer potential for personalized cancer therapy.
More Related Videos
09:34A Combinatorial Single-cell Approach to Characterize the Molecular and Immunophenotypic Heterogeneity of Human Stem and Progenitor Populations
Published on: October 25, 2018
10:23Single-cell Gene Expression Using Multiplex RT-qPCR to Characterize Heterogeneity of Rare Lymphoid Populations
Published on: January 19, 2017