Boolean Feedforward Neural Network Modeling of Molecular Regulatory Networks for Cellular State Conversion

Area of Science:

• Systems Biology
• Computational Biology
• Genomics

Background:

• Molecular regulatory networks (MRNs) govern cellular states and transitions.
• Modeling MRNs is essential but hindered by the difficulty of obtaining time-series experimental data.
• Single-cell RNA-sequencing offers pseudo-time ordering of cells, providing a novel data source.

Purpose of the Study:

• To propose a novel Boolean feedforward neural network (FFN) modeling approach.
• To reconstruct practical and useful MRN models from large temporal datasets.
• To identify control targets for cellular state conversion through MRN analysis.

Main Methods:

• Combined neural network and Boolean network modeling.
• Utilized single-cell RNA-sequencing data for pseudo-time ordering.
• Developed a parameter-free, logical modeling approach robust to noisy data.

Main Results:

• Successfully reconstructed MRN models using the Boolean FFN approach.
• Demonstrated the model's applicability on a toy model and biological networks.
• Identified potential control targets for cellular state conversion.

Conclusions:

• Boolean FFN modeling provides a robust method for MRN reconstruction from large temporal datasets.
• This approach facilitates the identification of key regulatory targets for manipulating cellular states.
• The method overcomes limitations associated with traditional time-series experimental data acquisition.