CCPE: Cell Cycle Pseudotime Estimation for Single Cell RNA-seq Data

Abstract

The rapid development of single-cell RNA-sequencing (scRNA-seq) technologies makes it possible to characterize cellular heterogeneity by detecting and quantifying transcriptional changes at the single-cell level. Pseudotime analysis enables to characterize the continuous progression of various biological processes, such as cell cycle. Cell cycle plays an important regulatory role in cell fate decisions and differentiation and is also often regarded as a confounder in scRNA-seq data analysis when analyzing the role of other factors on transcriptional regulation. Therefore, accurate prediction of cell cycle pseudotime and identify cell stages are important steps for characterizing the development-related biological processes, identifying important regulatory molecules and promoting the analysis of transcriptional heterogeneity. Here, we develop CCPE, a novel cell cycle pseudotime estimation method to characterize cell cycle timing and determine cell cycle phases from single-cell RNA-seq data. CCPE uses a discriminative helix to characterize the circular process and estimates pseudotime in the cell cycle. We evaluated the model performance based on a variety of simulated and real scRNA-seq datasets. Our results indicate that CCPE is an effective method for cell cycle estimation and competitive in various downstream analyses compared with other existing methods. CCPE successfully identified cell cycle marker genes and is robust to dropout events in scRNA-seq data. CCPE also has excellent performance on small datasets with fewer genes or cells. Accurate prediction of the cell cycle in CCPE effectively contributes to cell cycle effect removal across cell types or conditions.