Scaffolding

Scaffolds provide the necessary structural support and microenvironment for cells to grow and organize into functional 3D tissues. Without scaffolds, the cells may grow but will not form a useful mass that can be formulated into a product. The primary challenge lies in designing scaffolds that fit the purpose of the production, such as mimicking natural muscle tissue in the production of cultivated meat. Innovation on this key problem can be sped up through the use of artificial intelligence.

Two studies by Andrews et al. (2023, 2025) address this challenge by integrating biophysical simulations, deep learning, and evolutionary strategies to optimize mould designs for lab-grown tissues. Andrews et al. (2023) introduce a deep learning model to rapidly predict tissue properties like cell organization and alignment within moulds. In their later work, Andrews et al. (2025) employ evolutionary algorithms to efficiently explore and refine mould shapes, optimizing for cellular alignment and uniform density. These are the preliminary steps in the direction of identifying moulds that support and promote cellular growth toward a desired outcome.

Complementing these efforts, Rafieyan et al. (2024) and Bermejillo Barrera et al. (2021) focus on predicting the quality and mechanical properties of 3D printed scaffolds using various machine learning techniques. Bermejillo Barrera et al. (2021) utilize 3D convolutional neural networks to predict critical mechanical and textural properties. Rafieyan et al. (2024) developed a comprehensive dataset and deep learning algorithms to predict the overall quality of 3D bioprinted scaffolds based on material composition and printing parameters. Together, these studies demonstrate how AI can be leveraged to predict and optimize scaffold characteristics, removing one further barrier to producing final products through cellular agriculture.