AI Unveils Embryonic Development Stages with Precision

Animal embryos undergo a series of distinct developmental stages as they progress from a fertilized egg cell to a fully functional organism. This biological process is primarily controlled by genetics and follows a similar pattern across various animal species. However, there are variations in the specifics, both between different species and among embryos of the same species. For instance, the pace at which individual embryonic stages are completed can differ. These differences in embryonic development play a significant role in driving evolution, potentially leading to the emergence of new traits and thereby promoting evolutionary adaptations and biodiversity. Understanding the mechanisms of evolution greatly depends on studying the embryonic development of animals. To objectively and efficiently record differences in embryonic development, such as the timing of developmental stages, researchers at the University of Konstanz, led by systems biologist Patrick Müller, are developing and utilizing methods based on artificial intelligence (AI). In their recent article in Nature Methods, they outline a new approach that autonomously captures the pace of development processes and identifies characteristic stages without human intervention, standardizing the process across species boundaries.

Each embryo possesses unique characteristics

Our current understanding of animal embryogenesis and individual developmental stages stems from microscopic observations and detailed descriptions of embryos at various stages of development. Thanks to these meticulous efforts, reference books containing standardized depictions of individual embryonic stages are now available for numerous animal species.

However, embryos often do not look exactly the same under the microscope as they do in the schematic drawings. And the transitions between individual stages are not abrupt, but more gradual.”

Patrick Müller, systems biologist

Assigning an embryo to different stages of development is not easy, even for experts, and is somewhat subjective. Complicating matters further, embryonic development does not always proceed as expected. According to Müller, various factors like temperature can affect the timing of embryonic development. The AI-supported method developed by Müller and his colleagues represents a significant advancement. In an initial application, the researchers trained their Twin Network using over 3 million images of healthy developing zebrafish embryos. Subsequently, they utilized the resulting AI model to automatically determine the developmental age of other zebrafish embryos. The researchers successfully demonstrated that the AI can accurately identify key stages in zebrafish embryogenesis and automatically detect individual developmental stages without human intervention. In their study, the AI system was used to compare the developmental stage of embryos and describe the temperature dependency of embryonic development in zebrafish. Despite being trained with images of normally developing embryos, the AI could also identify abnormalities that may occur spontaneously in a certain percentage of embryos or be induced by environmental toxins.