SEMgraph: an R package for causal network inference of high-throughput data with structural equation models

Area of Science:

• Computational Biology
• Bioinformatics
• Systems Biology

Background:

• High-throughput sequencing generates vast, heterogeneous biological data.
• Integrating diverse data and existing knowledge is crucial for understanding complex biological systems and diseases.
• Scalable statistical modeling is essential for analyzing these large datasets.

Purpose of the Study:

• To develop a scalable statistical solution for modeling complex biological systems.
• To integrate heterogeneous data and existing knowledge for hypothesis testing.
• To improve comprehension of physiological processes and diseases using network analysis.

Main Methods:

• Developed the R package SEMgraph, combining network analysis and causal inference.
• Utilized structural equation modeling (SEM) framework.
• Implemented a fully automated toolkit for managing biological systems as multivariate networks.

Main Results:

• SEMgraph provides robust and reproducible data-driven evaluation of model architecture and perturbation.
• The package interprets results in terms of causal effects among system components.
• Offers a toolkit for analyzing complex biological networks.

Conclusions:

• SEMgraph facilitates the integration of heterogeneous data for causal inference in biological systems.
• The package enhances the understanding of physiological processes and diseases.
• Provides a valuable tool for researchers in molecular biology and medicine.