Parameter estimation of qualitative biological regulatory networks on high performance computing hardware

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Biological Regulatory Networks (BRNs) govern essential organismal functions through complex dynamics.
• Parameter estimation in BRNs is computationally intensive, especially for large networks.
• Existing sequential tools like SMBioNET face scalability challenges with increasing BRN complexity.

Purpose of the Study:

• To develop a computationally efficient method for parameter estimation in complex Biological Regulatory Networks.
• To leverage High Performance Computing (HPC) for accelerating the analysis of BRNs.
• To apply the enhanced method to identify therapeutic targets in cancer progression.

Main Methods:

• Implemented a parallel approach using data decomposition and MPJ Express (Java messaging library).
• Divided the parameter space into regions for parallel exploration on HPC hardware.
• Utilized a Java-based implementation for broad hardware compatibility.

Main Results:

• Achieved almost linear speed-up on multicore and cluster computers for various BRN sizes.
• Demonstrated application on the Hexosamine Biosynthetic Pathway (HBP) in cancer, identifying potential therapeutic targets.
• Evaluated performance on a 23-entity Fibroblast Growth Factor Signalling network in Drosophila melanogaster.

Conclusions:

• The parallel implementation offers significant computational speed-up for qualitative modeling of BRNs.
• The approach provides a scalable solution for parameter estimation on diverse computing platforms.
• The method facilitates the identification of biological system recovery states and therapeutic interventions.