DNFE: Directed network flow entropy for detecting tipping points during biological processes

Area of Science:

• Systems Biology
• Bioinformatics
• Genomics

Background:

• Dynamic biological processes exhibit critical tipping points marking state transitions.
• Traditional methods using undirected networks struggle with high-dimensional, small-sample omics data, especially single-cell data.
• Identifying tipping points and their driving networks is crucial for predicting and mitigating biological shifts.

Purpose of the Study:

• To develop a robust method for identifying tipping points and dynamic network biomarkers from omics data.
• To address limitations of traditional network analysis methods in high-dimensional biological datasets.
• To explore gene regulatory relationships and identify key drivers in biological transitions.

Main Methods:

• Developed the directed network flow entropy (DNFE) method to transform omics data into directed networks.
• Applied DNFE to diverse datasets including single-cell RNA-sequencing (scRNA-seq), bulk tumor, and blood data.
• Validated the method through numerical simulations on various noise levels and large-scale gene regulatory networks.

Main Results:

• DNFE effectively identified critical states and their dynamic network biomarkers across multiple real-world datasets.
• The method demonstrated robustness to noise and outperformed existing approaches in tipping point detection.
• DNFE successfully predicted active transcription factors and identified previously overlooked "dark genes".

Conclusions:

• The DNFE method provides a robust and effective framework for analyzing dynamic biological processes using omics data.
• DNFE enhances the identification of critical states, regulatory networks, and key genes, advancing systems biology research.
• This approach is applicable to both single-cell and bulk omics data, offering broad utility in biological discovery.