Mitigating Cell Cycle Effects in Multi-Omics Data: Solutions and Analytical Frameworks

Cell cycle structures vary significantly across cell types, which exhibit distinct phase compositions. Asynchronous DNA replication and dynamic cellular characteristics during the cell cycle result in considerable heterogeneity in DNA dosage, chromatin accessibility, methylation, and expression. Nonetheless, the consequences of cell cycle disruption in the interpretation of multi-omics data remain unclear. Here, we systematically assessed the influence of distinct cell phase structures on the interpretation of omics features in proliferating cells, and proposed solutions for each omics dataset. For copy number variation (CNV) calling, asynchronous replication timing (RT) interference induces false CNVs in cells with high S-phase ratio (SPR), which are significantly decreased following replication timing domain (RTD) correction. Similar noise is observed in the chromatin accessibility data. Moreover, for DNA methylation and transcriptomic analyses, cell cycle-sorted data outperformed direct comparison in elucidating the biological features of compared cells. Additionally, we established an integrated pipeline to identify differentially expressed genes (DEGs) after cell cycle phasing. Consequently, our study demonstrated extensive cell-cycle heterogeneity, warranting consideration in future studies involving cells with diverse cell-cycle structures. RTD correction or phase-specific comparison could reduce the influence of cell cycle composition on the analysis of the differences observed between stem and differentiated cells.