A method for evaluating the degree of Adipogenic differentiation of porcine cells cultured in suspension based on deep learning.
Journal:
Food research international (Ottawa, Ont.)
Published Date:
Jul 1, 2025
Abstract
Fat plays a very important role in the quality of meat and meat products. Cell culture based adipocyte differentiation is a crucial step in the production of cell-cultured meat, traditionally relying on quantitative assessments through fluorescence staining image analysis or molecular biology experiments. However, these methods are time-consuming, labor-intensive, and easily influenced by the observer's perception. Therefore, we require a faster, more intuitive, and accurate evaluation technique. By combining high-throughput technology with deep learning, we analyzed bright-field images captured during the porcine suspended adipocyte differentiation process. Based on our existing cell differentiation protocols, we rapidly collected high-throughput bright-field images and developed a deep learning model to assess the adipogenic differentiation degree in these images. Through the mutual validation of single-shake and multi-shake bottles, the model achieved good results (RMSE = 5.90, R = 0.8321). It can determine the degree of adipogenic differentiation of a sample by inoculating 12 wells in a 96-well plate, allowing for accurate discrimination between samples with different levels of differentiation. This method has enabled us to establish an application that integrates high-throughput screening and deep learning, and has assisted us in identifying the potential adipogenic differentiation-promoting capability of arachidonic acid (100 μM) in the rapid screening of edible fatty acids and proteins.
