75 3 Towards Automatization of Organoid Analysis: A Deep Learning Approach to Localize and Quantify Organoids Images INTRODUCTION New advances in the biomedical field have presented us with the need for new technical and analytical approaches. The field of organoids has experienced a massive upsurge in the past 15 years. Organoids are three dimensional mini-structures that in culture recapitulate some features and functions of the organ of origin. Ever since organoids were first described in 20091, the field has continued to expand rapidly. After that, many different types of organoids, recapitulating different organs, have been described2–6. These three-dimensional structures have become widely popular due to their ability to self-organize and proliferate while maintaining multiple biological functions7. Organoids can be used for multiple purposes, ranging from developmental studies to diseases modeling, as well as, drug screening 7,8. Since it is the three-dimensional environment that allows cells to organize and behave in a more physiologically relevant manner, it is understandable how important morphology and size are7. It is common for researchers to track organoid morphology and size as these are the first signs of health and disease. This is often done for multiple purposes: either for developmental reasons/optimization of the culture conditions9,10, or in the context of different diseases11. Another interesting approach to studying organoid health is tracking organoids’ size in time12. This is done in order to determine the growth and expansion rate of the culture in time. This proliferative status of the organoids can provide relevant information regarding developmental state of the organ, regeneration, response to pharmacological interventions. However, it is important to highlight that due to the way these structures are cultured, high variability in size is often observed between organoids in the same dome (Figure 1B and 1C). One of the main reasons for this is the way they are distributed in the hydrogel domes. Since these domes are three- dimensional structures (Figure 1A), not all organoids have access to the same amount of nutrients in the medium. Often, this causes organoids around the edges of the hydrogel to have easier access to nutrients, reaching bigger sizes than those organoids contained in the middle of the dome (Figure 1B and 1C). Another complexity arising from this type of culture is the proliferative state of the cells that leads to different organoids expanding and collapsing at different rates. Because of this high variability between organoids in the same dome, researchers need to quantify large numbers of organoids per well in order
RkJQdWJsaXNoZXIy MTk4NDMw