Comprehensive phenotyping of neuro-organoids by deep learning.

Identification of disease mechanisms and novel therapeutic targets relies on the use of cell culture and animal models. While the former are overly simplified, the latter are not human and ethically contested. Suboptimal models at the discovery side will inevitably lead to a steep loss of leads in clinical trials. With the advent of human induced pluripotent stem cell technology, it has now become possible to generate organoids that more faithfully capture part of the heterogeneity and three-dimensional context of human tissue. Several research labs at the University of Antwerp (UA) recognize their potential and have therefore implemented a variety of human patient-derived organoid cultures, in particular for neuroscientific research lines. However, batch-to-batch variability and the inability to characterize these specimens at the cellular level with high-throughput, hamper their integration in a routine screening setting. Therefore, we have the ambition to develop an end-to-end solution that enables unbiased cellular phenotyping of intact neuro-organoids by using a combination of fluorescent labelling, advanced microscopy, and artificial intelligence (AI).

Keywords:MULTISCALE IMAGE ANALYSIS, LIGHT SHEET MICROSCOPY, DEEP LEARNING, ORGANOID

Disciplines:Cell, tissue and organ engineering, Computational biomodelling and machine learning, Data visualisation and high-throughput image analysis, Neurological and neuromuscular diseases