Benchmarking scRNA-seq Copy Number Inference: A Comprehensive Evaluation and Practitioner's Guide.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate copy number variation (CNV) inference from single-cell RNA sequencing (scRNA-seq) is crucial for cancer research.
• Existing CNV tools show variable performance, lacking comprehensive benchmarking for reliable guidance.

Purpose of the Study:

• To conduct a broad benchmark of 12 CNV inference methods using diverse scRNA-seq datasets.
• To provide a practical decision-making framework for selecting appropriate CNV inference tools.

Main Methods:

• Evaluated 12 CNV inference methods across 28 real-world scRNA-seq datasets (>100,000 cells) and synthetic datasets.
• Assessed methods based on malignant cell classification accuracy, CNV inference accuracy, scalability, and robustness.

Main Results:

• Established a definitive guideline for CNV inference tool selection.
• Allele-aware methods (e.g., Numbat) perform best with high-quality allelic data.
• Expression-centric tools (e.g., Clonalscope, CopyKAT, inferCNV, SCEVAN) are reliable when raw sequencing data is unavailable.

Conclusions:

• The study provides a practical framework to guide researchers in choosing CNV inference methods.
• A public repository of standardized CNV profiles was created to foster methodological advancements in cancer genomics.