InferPloidy: A fast ploidy inference tool accurately classifies cells with abnormal CNVs in large single-cell RNA-seq datasets

Estimation of copy number variation (CNV) and ploidy inference provides valuable insights into the structural and clonal characteristics of tumor cells in single-cell RNA-seq datasets. Several tools have been developed for these purposes, including CopyKat and SCEVAN, but one limitation is their running times, which hinder their use for large datasets. In this study, we present InferPloidy, a faster and more accurate ploidy inference tool that operates on top of InferCNV. Rather than focusing on precise CNV segmentation or the discovery of intra-tumoral heterogeneity, InferPloidy emphasizes much faster and more accurate classification of aneuploid cells from diploids, which is crucial for identifying diagnostic markers or druggable targets. The accurate identification of malignant cells and the scalability of the tool to handle large datasets with many samples are key aspects of this work. InferPloidy is two orders of magnitude faster than existing tools while providing more accurate classifications, enabling fast and precise downstream analyses, including inter-tumoral heterogeneity studies across different patients.