Reconstructing Dynamic Gene Regulatory Networks Using f-Divergence from Time-Series scRNA-Seq Data.

Area of Science:

• Computational Biology
• Systems Biology
• Genomics

Background:

• Inferring dynamic gene regulatory networks (GRNs) from time-series single-cell RNA sequencing (scRNA-seq) data is complex.
• Existing methods struggle with scRNA-seq data's sparsity, dropouts, and heterogeneity, and often fail to capture dynamic regulatory changes at the single-cell level.

Purpose of the Study:

• To propose a novel method, f-divergence-based dynamic gene regulatory network inference (f-DyGRN), for reconstructing time-varying GRNs from scRNA-seq data.
• To address the limitations of current methods in capturing dynamic regulatory changes in single cells over time.

Main Methods:

• Utilized f-divergence to quantify temporal gene expression variations in individual cells.
• Integrated a first-order Granger causality model with regularization and partial correlation analysis.
• Employed a moving window strategy to capture dynamic gene interactions across different time stages.

Main Results:

• f-DyGRN demonstrated superior performance in reconstructing dynamic regulatory networks compared to existing methods.
• The method successfully analyzed both simulated and real scRNA-seq data from THP-1 cells.
• Performance was dependent on the choice of the f-divergence measure.

Conclusions:

• f-DyGRN offers a robust approach for inferring dynamic gene regulatory networks from time-series scRNA-seq data.
• The method effectively handles the unique challenges posed by scRNA-seq data.
• This advancement aids in understanding dynamic biological processes at the single-cell level.