Improving gene network inference with graph wavelets and making insights about ageing-associated regulatory changes...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell expression profiles offer detailed biological insights but are susceptible to noise and batch effects.
• Accurate gene regulatory network (GRN) inference is crucial for understanding cellular responses to various factors.

Purpose of the Study:

• To develop a robust method for improving gene network analysis in single-cell transcriptomics.
• To enhance the accuracy and consistency of GRN inference, particularly in the presence of technical noise and batch effects.

Main Methods:

• A novel graph-wavelet filter-based approach named GWNet was devised.
• GWNet was integrated with existing gene network inference methods to assess its performance.
• The method's ability to handle sparse single-cell expression data and batch effects was evaluated.

Main Results:

• GWNet significantly improved the performance of multiple gene network inference algorithms.
• The method demonstrated enhanced consistency in predicting gene regulatory networks from single-cell transcriptomes, even with batch effects.
• GWNet enabled reliable identification of gene regulatory changes missed by differential-expression analysis.

Conclusions:

• GWNet provides a robust framework for analyzing gene regulatory networks in single-cell transcriptomics.
• The approach facilitates the discovery of biologically relevant regulatory changes, such as those associated with aging in lung cells.
• Age-related regulatory patterns in lung cells show striking similarities to those induced by novel coronavirus infection.