Fig. 2: Morphology assessment of brain organoids identifies Feret diameter as a classifier of organoid quality.
From: Deciphering brain organoid heterogeneity by identifying key quality determinants
A Schematic illustrating the strategy to assess and determine morphological parameters that align with expert evaluation. B Bar graph depicting the result of the correlation analysis of morphological features with the expert’s quality evaluation. The red square highlights five parameters surpassing the stringency threshold of p-value < 10−5. C Venn diagram depicting the overlap of organoids evaluated by an expert (green and pink, High, Low) or clustered in an unbiased way into two groups with k-means clustering using five highly correlated parameters (light-green and dark-pink, ‘Cluster 1’ and ‘Cluster 2’). Positive (PPV) and negative prediction values (NPV) of the k-means clustering are depicted in the graph. D Bar graph illustrating Youden indices showing the diagnostic properties of five morphological parameters. The number on top of the graph represents the index value. E ROC curve for Feret diameter illustrating the ideal threshold with maximized specificity and sensitivity. Identified Feret diameter threshold: 3050 μm. F Venn diagram depicting the overlap of organoids evaluated by an expert (green and pink, High, Low) or clustered in two groups using Feret diameter threshold (light-green and dark-pink, Feret Low - Low Feret diameter 1’ and Feret High - High Feret diameter 2’). Positive (PPV) and negative prediction values (NPV) of the k-means clustering are depicted in the graph.
