LogicSR: prior-guided symbolic regression for gene regulatory network inference from single-cell transcriptomics...

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Deciphering gene regulatory mechanisms from high-dimensional single-cell data is challenging due to data sparsity, noise, and complex transcription factor (TF)-mediated regulation.
• Existing methods struggle to accurately model dynamic combinatorial regulatory logic inherent in biological systems.

Purpose of the Study:

• To introduce LogicSR, a novel computational framework for reconstructing gene regulatory networks (GRNs) from single-cell gene expression data.
• To improve the accuracy and interpretability of GRN inference by integrating Boolean logic models with symbolic regression.

Main Methods:

• LogicSR integrates mechanistic interpretability of Boolean logical models with equation-discovery capabilities of symbolic regression.
• A multi-objective Monte Carlo tree search (MCTS) framework incorporates prior biological knowledge to ensure plausibility and accelerate the search for governing equations.
• The framework was evaluated on synthetic and real-world benchmark datasets, including human embryonic stem cell data.

Main Results:

• LogicSR demonstrates high accuracy in reconstructing gene regulatory networks from single-cell data.
• The framework outperforms existing methods on both synthetic and real-world benchmark datasets.
• Application to human embryonic stem cell data successfully elucidated complex combinatorial TF-target gene regulations and identified key regulators.

Conclusions:

• LogicSR provides a powerful and accurate computational framework for deciphering gene regulatory mechanisms from single-cell data.
• The integration of Boolean logic and symbolic regression offers a promising approach for understanding complex biological regulation.
• LogicSR has significant potential for advancing systems biology research and identifying critical regulators in various biological contexts.