Fig. 4: Identification of mutations that have an impact in the morphology of an individual neuron.
From: CAJAL enables analysis and integration of single-cell morphological data using metric geometry
a Schematic of approach for identifying features (gene expression, mutations, protein expression, etc.) associated with cell morphological changes based on multi-modal data. For each feature, the degree of consistency between the feature values and the structure of the cell morphology space is quantified using the Laplacian score (C). Features with a low score are associated with local regions of the cell morphology space. The statistical significance of each feature in relation to the covariates is evaluated by means of a one-sided permutation test. In the figure, examples of features that are significantly localized in the cell morphology space (feature 1, a small number of random configurations have a smaller value of Cfeature, independently of the value of Ccovariate), not significantly localized in the cell morphology space (feature 2, a large number of random configurations have a smaller value of Cfeature), and substantially localized in the morphology space but in association with the covariate (feature 3, a small number of random configurations have smaller value of Cfeature, but they are not independent on the value of Ccovariate), are presented. b Mutations that have an impact on the morphology of the DVB interneuron in C. elegans. Null alleles are ranked according to their Laplacian score (C) in the cell morphology space of the DVB interneuron. The age of the worm was used as a covariate. Genes that significantly impact the morphology of the DVB interneuron are indicated in red (FDR < 0.05). c, d UMAP visualization of the cell morphology space of the DVB interneuron colored by the age of each worm (c) and the mutation status of unc-97, nlg-1, nrx-1, and unc-25 (d) (red: mutated; gray: wild-type). e Restricting the analysis to worms of the same age allows us to identify the age of onset of the morphological effects induced by each significant mutation (FDR < 0.05). Dashed lines indicate time points for which there is limited data to restrict the analysis. Source data are provided as a Source data file.
