Fig. 1

A Schematic shows that while variants in multiple candidate genes each contribute a small effect size towards cases of neurodevelopmental disorders, they may each affect overlapping molecular functions, providing a shared etiology for distinct subtypes of developmental disorders. B Violin plots show the average connectivity of 68 cytoplasmic stress granule genes, annotated from the Gene Ontology database (GO:0010494), with candidate neurodevelopmental genes from disorder-specific databases (DBD: https://dbd.geisingeradmi.org/; DDDG2P: https://www.ebi.ac.uk/gene2phenotype/; SFARI: https://gene.sfari.org/; Schizophrenia [6]; Epilepsy [7], in the context of a human brain-specific interaction network. Average connectivity was calculated as the shortest distance between two genes in the network, which was previously constructed using a Bayesian classifier trained on brain co-expression datasets [8]. * indicates gene categories with significantly less connectivity than the average connectivity of stress granule genes with all genes in the genome (p < 0.05, two-tailed paired t tests)