LogBTF: gene regulatory network inference using Boolean threshold network model from single-cell gene expression...

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Inferring gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNA-seq) data is crucial for understanding cellular mechanisms.
• Existing GRN inference methods often focus on network topology and neglect the dynamics of gene regulation.
• Noise and overfitting in time-series data pose significant challenges for accurate GRN inference.

Purpose of the Study:

• To develop a novel method, LogBTF, for inferring GRNs that explicitly models regulatory logic and dynamics.
• To address the limitations of existing methods in handling time-series data noise and overfitting.
• To improve the accuracy of GRN inference from scRNA-seq data.

Main Methods:

• Proposed LogBTF, an embedded Boolean threshold network method integrating regularized logistic regression and Boolean threshold functions.
• Converted continuous gene expression data to Boolean values and used elastic net regression.
• Implemented a perturbation design matrix and cross-validation to optimize network topology and mitigate overfitting and multicollinearity.

Main Results:

• LogBTF effectively infers GRNs by integrating binarized time-series data with logistic regression and Boolean threshold functions.
• The method successfully overcomes multicollinearity and overfitting issues through a novel perturbation approach.
• Extensive experiments on simulated and real scRNA-seq datasets demonstrated superior accuracy of LogBTF compared to alternative methods.

Conclusions:

• LogBTF provides a robust and accurate approach for inferring gene regulatory network dynamics from time-series single-cell RNA sequencing data.
• The method's ability to handle noise and overfitting enhances its utility in complex biological systems.
• LogBTF represents a significant advancement in computational methods for systems biology research.