Using single cell atlas data to reconstruct regulatory networks.

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Inferring gene regulatory networks (GRNs) from omics data is crucial in systems biology.
• Existing methods for transcription factor (TF)-gene interaction inference often use limited or static data, hindering the analysis of dynamic biological processes.
• Temporal dynamics are essential for understanding gene regulation.

Purpose of the Study:

• To develop an advanced computational method for inferring global gene regulatory networks.
• To overcome limitations of existing methods by incorporating temporal information.
• To achieve more accurate and comprehensive identification of regulatory interactions.

Main Methods:

• A novel computational method combining neural networks and multi-task learning was developed.
• The method predicts RNA velocity, a measure of dynamic gene expression changes, instead of static gene expression values.
• Application involved analyzing atlas-scale single-cell data from 6 Human BioMolecular Atlas Project (HuBMAP) tissues.

Main Results:

• The new method demonstrated superior performance compared to prior approaches for GRN inference.
• It identified a more comprehensive set of regulatory interactions.
• Validated and novel TF-gene interaction predictions were generated from the HuBMAP dataset.

Conclusions:

• The developed RNA velocity-based method significantly enhances the accuracy and scope of gene regulatory network inference.
• This approach is well-suited for analyzing inherently temporal biological processes using large-scale single-cell data.
• The findings provide a more robust tool for systems biology research and understanding complex regulatory mechanisms.