Predicting heterogeneity in clone-specific therapeutic vulnerabilities using single-cell transcriptomic signatures.

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Precision oncology increasingly recognizes the importance of intra-tumor heterogeneity (ITTH) alongside inter-tumor heterogeneity.
• Understanding ITTH is crucial for predicting patient clinical outcomes and tailoring cancer treatments.

Purpose of the Study:

• To highlight the significance of intra-tumor transcriptomic heterogeneity (ITTH) in predicting clinical outcomes.
• To develop and validate a computational system (CaDRReS-Sc) for predicting drug response based on ITTH using single-cell RNA sequencing (scRNA-seq).

Main Methods:

• Large-scale analysis of cancer omics datasets to identify the importance of ITTH.
• Leveraging single-cell RNA sequencing (scRNA-seq) data with a recommender system (CaDRReS-Sc).
• Validation using patient-proximal cell lines for monotherapy and combinatorial drug predictions.

Main Results:

• Heterogeneous gene-expression signatures derived from scRNA-seq can predict drug response with 80% accuracy.
• CaDRReS-Sc demonstrated high accuracy in predicting monotherapy responses (Pearson r > 0.6) and combinatorial effects (>10% improvement).
• The system effectively targets clone-specific vulnerabilities within tumors.

Conclusions:

• Intra-tumor transcriptomic heterogeneity is a critical factor in predicting cancer patient outcomes and drug responses.
• CaDRReS-Sc provides a powerful in silico tool for predicting therapeutic vulnerabilities and accelerating drug repurposing.
• The system's application to scRNA-seq compendiums offers a scalable approach for cancer research and drug discovery.