LASSIM-A network inference toolbox for genome-wide mechanistic modeling

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Advancements in multi-omics data generation enable complex biological system analysis.
• Systems pharmacology requires integration of diverse biological datasets.
• Gene regulatory networks (GRNs) are crucial for understanding cellular processes.

Purpose of the Study:

• Introduce Large-scale simulation modeling (LASSIM), a novel mathematical tool for GRN inference.
• Enable large-scale inference using mechanistically defined ordinary differential equations (ODEs).
• Integrate structural knowledge of regulatory interactions with multi-omics data.

Main Methods:

• LASSIM utilizes a two-step approach: inferring ODE systems for core genes and optimizing parameters for peripheral gene regulation.
• Implemented as a general-purpose toolbox using the PyGMO Python package for parallel computation.
• Leverages multicore computers and high-performance clusters for efficient analysis.

Main Results:

• Successfully inferred a large-scale non-linear model of naive Th2 cell differentiation.
• Integrated Th2-specific bindings, time-series data, and siRNA-mediated knockdown experiments.
• Demonstrated superior performance compared to existing models in monitoring T-cell differentiation data.

Conclusions:

• LASSIM toolbox facilitates a new paradigm in model-based data analysis, merging mechanistic models with systems-level data.
• Enables the inference of genome-wide transcription regulatory systems.
• Provides a powerful approach for studying systems involved in immune-related diseases.