A new asynchronous parallel algorithm for inferring large-scale gene regulatory networks.

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Reconstructing gene regulatory networks (GRNs) is crucial in systems biology.
• Existing algorithms struggle with accuracy and computational cost for large-scale GRNs.
• High-throughput data and complex biological problems necessitate scalable inference methods.

Purpose of the Study:

• To design an effective parallel algorithm for inferring large-scale GRNs.
• To improve accuracy and reduce time complexity in GRN reconstruction.
• To leverage high-performance parallel computing environments for biological network analysis.

Main Methods:

• Proposed a novel asynchronous parallel framework combining splitting technology and ordinary differential equation (ODE)-based optimization.
• Utilized GRN sparsity and modularity to decompose large networks into smaller subnetworks.
• Performed parallel ODE-based optimization and asynchronous communication for network parameter inference.

Main Results:

• The parallel algorithm demonstrated superior performance in inferring large-scale GRNs.
• Achieved higher accuracy and lower time complexity compared to existing popular algorithms.
• Validated on benchmark datasets (DREAM), E. coli GRN, and a dataset with over 10,000 genes.

Conclusions:

• The developed asynchronous parallel framework effectively addresses challenges in large-scale GRN inference.
• The approach offers a significant improvement in both accuracy and computational efficiency.
• This method provides a scalable solution for systems biology research using high-performance computing.