Design optimization of geometrically confined cardiac organoids enabled by machine learning techniques

Area of Science:

• Biomedical Engineering
• Stem Cell Biology
• Computational Biology

Background:

• Stem cell organoids are valuable for studying organ development, disease, and drug screening.
• Integrating organoid technology with tissue engineering and AI can advance organoid design principles.

Purpose of the Study:

• To explore structure-function relationships in cardiac organoids using machine learning.
• To develop optimized cardiac organoid designs through geometric manipulation and data analysis.

Main Methods:

• Fabricated 230 cardiac organoids with 7 distinct geometric designs using micropatterning.
• Applied manifold learning to analyze organoid heterogeneity based on 10 physiological parameters.
• Utilized unsupervised machine learning for clustering and functional similarity assessment.

Main Results:

• Identified unique functionalities associated with specific geometric designs in cardiac organoids.
• Demonstrated the critical role of calcium transient rising time in differentiating organoids.
• Clustered organoids based on functional similarity, linking geometry to performance.

Conclusions:

• Machine learning integration enhances understanding of cardiac organoid structure-function relationships.
• Geometric design significantly influences cardiac organoid functionality.
• This approach enables more controlled and optimized organoid engineering for future applications.