Fig. 6: Molecular analyses in mice and humans place NEUROD2 as a nexus in NDD gene regulatory network.

a Scheme of the RNA-seq experiment, performed at P30. b Gene set enrichment analysis using the DAVID knowledgebase. c Fold change expression (FC; log2 scale) of 13 DEX genes belonging to the voltage-dependent ion channel family, ranked according to Padj (lowest on top). Genes associated with neuropsychiatric recurrent syndromes are depicted in blue. d Heatmap for synaptic and disease gene sets among the orthologs of DEX genes (red, higher expression; blue, lower expression). e, f Venn diagrams identifying overlaps between DEX genes and synaptic/ NDD gene sets. Number of genes for each gene set is indicated. Right in (f), differentially-expressed genes associated with ASD and their score in SFARI. g Graphical representation of % [DEX genes] (orange) and % [same number of randomly selected genes] (gray) belonging to gene sets of interest. (h) qRT-PCR for 11 neuropsychiatric-related DEX genes in cortical samples from WT, HET, and KO. These data validate the RNA-seq results and show that HET mice have an intermediate phenotype between WT and KO mice, demonstrating that Neurod2 is haploinsufficient. (i–k) NEUROD2-based coexpression network analysis in humans. i Gene module 37 with spatio-temporal co-expression was obtained from WGCNA of spatiotemporal transcriptional dynamic analyzes in the human cortex [49]. j Network representation of NEUROD2 within module 37 from Li et al. [37] showing gene connectivity based on Pearson correlation (weight of edges scaled for r between 0.7 and 1). The size of each node is proportional to degree centrality calculated with a cutoff of r = 0.7. Green edges indicate direct neighbors of NEUROD2 with threshold r = 0.7. Genes described in SFARI are colored according from SFARI score (1–3). k Circular network representation of NEUROD2 and SFARI genes interactions with weight of edges scaled for r between 0.5 and 1. The size of each node represents the same values of degree centrality as calculated in (j). For (j, k), we added a circle with nine cells. Each cell corresponds to a neuropsychiatric trait/disorder and color/position legend is at the bottom. For details on how neuropsychiatric trait/disorder cells were determined, see “Materials and methods” and File S1, sheet “GWAS_NEUROD2_GeneModule_Fig6”. N = 3 experiments per genotype for both RNA-seq and qRT-PCR. Data are means ± SEM. Statistical significance was evaluated by binomial test (g) or unpaired two-tailed Student t test (h) (*P < 0.05; **P < 0.01; ****P < 0.0001). See also Figs. S11 and S12.