Toward synthesizing executable models in biology.

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Executable models simulate biological behaviors, aiding scientific discovery and guiding experiments.
• Formal methods like model checking enhance executable model analysis for comprehensive validation.
• Automated program synthesis offers a pathway to generate executable models from specifications.

Purpose of the Study:

• To survey recent advancements in applying computer program synthesis techniques to infer biological models from experimental data.
• To highlight the potential of synthesis for automated biological model construction.
• To explore new opportunities offered by synthesis in biological data analysis.

Main Methods:

• Surveying techniques underlying computer program synthesis for biological model inference.
• Describing synthesis of biological models from curated mutation data.
• Detailing inference of network connectivity from phosphoproteomic data.
• Explaining synthesis of Boolean networks from gene expression data.

Main Results:

• Synthesis techniques can automatically extract executable biological models from diverse experimental datasets.
• This approach enables efficient computation of disambiguating experiments.
• Automated generation of various model types from biological data is feasible.

Conclusions:

• Synthesis offers a powerful, automated approach to biological model inference, complementing traditional machine learning methods.
• This methodology facilitates deeper understanding of biological systems and accelerates research.
• Future applications include automated model discovery and experimental design for biological systems.