Design of a flexible component gathering algorithm for converting cell-based models to graph representations for use...

Area of Science:

• Computational Biology
• Regenerative Biology
• Systems Biology

Background:

• The rapid generation of experimental data in science, particularly in regenerative biology, has outpaced the ability to integrate it into conceptual frameworks for higher-level understanding.
• Multidimensional and shape-based data in regenerative biology, such as that for planaria, presents a significant challenge for understanding body form maintenance.
• PlanformDB, a novel repository, stores planarian experiment descriptions and morphological outcomes using graph formalism, offering a data resource for computational modeling.

Purpose of the Study:

• To develop an automated model discovery framework for regenerative biology.
• To identify plausible biological mechanisms underlying planarian regeneration using computational approaches.
• To bridge the gap between large-scale experimental data and mechanistic understanding in planarian biology.

Main Methods:

• A cell-based modeling platform integrated with evolutionary search was employed for automated model discovery.
• A flexible connected component algorithm converted cell-based simulation data into graph representations of virtual planaria.
• Graph-edit distance was utilized to quantitatively compare simulated morphology with experimental data from PlanformDB, guiding the evolutionary search.

Main Results:

• A cell-based model of planaria capable of regenerating anatomical regions after bisection was successfully developed.
• The automated model discovery framework demonstrated its capability to search for and identify models of planarian regeneration that align with experimental data.
• Quantitative metrics derived from graph-edit distance enabled the validation and refinement of computational models against empirical observations.

Conclusions:

• The developed algorithm for converting cell-based models to graphs facilitates automated development, training, and validation of computational models using morphology-based data.
• This work represents a significant step towards automating the search for biological mechanisms in regenerative biology.
• The automated framework promises to expand the capacity for identifying, considering, and testing hypotheses in the study of regeneration.