Automated Morphometric Analysis and Machine Learning Classification of iPSC-Derived Vascular Organoids in Alzheimer’s and Parkinson’s Disease Models
Roman V. Shumilin1, Anna A. Kopylova1, Daria D. Volegova1, Elizaveta S. Pereplitsa2, Anna V. Blagova2, Ksenia O. Salina1, Ilya K. Lanikin1, Petr I. Kupriyanov1, Anna V. Zubova2, Marina R. Kapkaeva2, Ivan V. Simkin1, Alla B. Salmina1,2, Stanislav O. Yurchenko1 and Sergei N. Illarioshkin2
1 Laboratory of Cellular Technologies and Tissue Engineering, Center “Soft Matter and Physics of Fluids,” Bauman Moscow State Technical University, Moscow, Russia
2 Laboratory of Neurobiology and Tissue Engineering, Brain Science Institute, Russian Center of Neurology and Neurosciences, Moscow, Russia
Abstract
Three-dimensional vascular organoids derived from induced pluripotent stem cells (iPSCs) provide a relevant in vitro model for studying morphological alterations associated with neurodegenerative diseases. This study focused on the development of an automated approach for morphometric analysis and classification of vascular organoids obtained from healthy donors and patients with Alzheimer’s and Parkinson’s diseases. Bright-field time-lapse images were acquired using a JuLI Stage inverted microscope at 4× magnification at 30-minute intervals. The dataset comprised 24632 images of vascular organoids. Image preprocessing included contrast enhancement using CLAHE, Butterworth filtering, adaptive thresholding, noise removal, contour detection, and subsequent contour processing.
The developed Python-based software enabled automated extraction of 15 morphometric parameters characterizing organoid size, shape, and contour properties. The extracted morphometric features were subsequently used as input parameters for machine-learning models designed to distinguish control vascular organoids from pathological groups. Boosting-based classifiers were developed for the automated identification of disease-associated morphological patterns. The highest classification performance was achieved for the discrimination between the control group and Parkinson’s disease, with an F1-score of 0.86. Overall, the proposed workflow integrates bright-field time-lapse microscopy, automated image processing, quantitative morphometric analysis, and machine learning within a unified computational framework for the characterization and classification of iPSC-derived vascular organoids.
This work was funded by a grant from the Ministry of Science and Higher Education of the Russian Federation, provided as subsidies from the federal budget for major scientific projects in priority areas of scientific and technological development (Agreement No. 075-15-2024-638 dated July 12, 2024)
Speaker
Roman Shumilin
Bauman Moscow State Technical University, Moscow, Russia
Russia
Discussion
Ask question
