Fig. 2: Global organisational heterogeneity reduces to two organisational archetypes.

a, b Organisational diversity increases over time. A Gaussian Mixture Model (GMM) stratifies organisational feature space into eleven ‘organisational motifs’ using pooled data from four developmental stages. Some motifs persist throughout early development (e.g. III, VI, IX), while others emerge dynamically (e.g. I, II, VII). Feature values are z-scored; clustering is performed across all time points. b Example motifs: Motif I is absent early and emerges from 24 hpf on, ultimately covering the CNS; Motif IX is consistently found in the YSL and later includes the median fin fold. Feature space (t-SNE) based on all time points. n: number of cells per motif in the shown sample. c–e Organisational diversity converges into two archetypes. c Hierarchical agglomerative clustering (HAC) of organisational motifs (48 hpf data) identifies two archetypes with opposing features. ‘Amorphous’ cells are dense, isotropic, and irregular; ‘crystalline’ cells are low-density, anisotropic, and regular. Feature values are z-scored. HAC used Euclidean distance and Ward linkage. d Archetypes map to cohesive spatial domains at 48 hpf. CNS, sensory organs, and fin buds are amorphous; muscles, skin, and YSL are crystalline. e Cells expressing tissue-specific markers align with one of the two archetypes, with the exception of fibroblasts. n mean number of cells per motif across N = 34 samples ± standard deviation. All scale bars 500 µm. Source data are provided as a Source Data file.