<p>Computational approach to identify clusters of cellular processes.</p>

Both approaches were used in parallel in our work. <b>(B)</b> Violin Plot showing each cluster from the larval dataset. Red bars below indicate whether each cluster was identified as processes by Approach 1, Approach 2, and if the cluster was excluded from the final atlas. <b>(C)</b> Violin Plot showing each subcluster from the adult dataset after over-clustering.

Red bars below indicate whether each subcluster was identified as processes by Approach 1, Approach 2 and if the subcluster was excluded from the final atlas. It also shows which original ID cluster the subclusters come from. <b>(D, E)</b> UMAP plots showing the original ID (D) and the over-clustering (E) of the adult dataset. The code used to generate the graphs can be found at <a href="zenodo.org/records/21890456" target="_blank">zenodo.org/records/21890456</a>.</p> <p>(TIFF)</p>