Alterations in signaling and lineage-specific gene regulatory networks (GRNs) in R-loop-depleted gastruloids. (A) The heatmap of communication probability of retinoic acid (RA) signaling among mesoderm cells (PSM-A, PSM-B, SM) and ectoderm cells (NMP1, NMP2, spinal cord). The sources and targets of inferred RA signaling are indicated. (B) The communication probabilities of indicated ligand-receptor pairs in RA signaling from PSM-A to mesoderm cell types (PSM-A, PSM-B, SM) and ectodermal cell types (NMP1, NMP2, spinal cord). (C) Barplots depicting the proportion of targets in the indicated module related to the total target in each specific cell type/treatment combination. (D) Network plots illustrating TF–target relationships in the Rarb module for spinal cord. The trajectory diagram highlights the cell type in the trajectory map. “Unique” edges indicate TF–target pairs that only exist in a specific combination of cell type and treatment, shown with a dashed line, whereas solid lines denote nonunique pairs. Edge thickness indicates the correlation strength of the TF–target pair. The line color of each edge indicates a positive or negative correlation between TF and its targets. (E) Dotplot of GO analysis results for target genes in the Rarb module. Shared targets indicate genes are positively correlated and shared between control and Dox. A Q-value ≤ 0.05 was used as the cutoff for GO category enrichment. (F) Network plots illustrating TF–target relationships in the Hand1, Hand2, and Mef2c modules for splanchnic mesoderm. Labels are as in D. (G) Dotplot of GO analysis results for target gene categories in Mef2c module. The positively correlated, control-specific targets (left) and the targets shared between the control and Dox (right) are shown. A Q-value ≤ 0.05 was used as the cutoff for GO category enrichment.