Fig. 4

Modality analysis of the CalMorph morphological parameters. A Mixture-model-based clustering was used to check the modality of 501 CalMorph parameters according to the data types of (i) non-negative, (ii) ratio, (iii) noise, and (iv) proportion measures. Using 114 replicates of wild-type yeast cells revealed 431 unimodal and 70 multimodal parameters. B Confounding factors due to the differences in the experimental conditions and outliers were further considered as multimodal behaviors. A GLM was introduced by constructing a linear model (one-way analysis of variance) of the confounding factors for each parameter. Outlier data points were removed after defining a threshold for each parameter (one-percentile rule). As a result, 490 unimodal parameters were detected