Fig. 2

Overlap between NMD and FMRP targets in neuronal cells or tissues, and subcellular localization of NMD and FMRP targets in neuronal cells. a Venn diagram showing the overlap between the 1027 NMD targets defined using mouse N2A neuroblastoma cells [33], and the 5733 FMRP targets defined using total mouse cerebral cortex [77], mouse hippocampus [77], mouse cerebellum [77], mouse hippocampal CA1 pyramidal neurons [78], mouse cerebellar granule cells [78], or mouse whole brain polysome fractions [79]. The statistical significance of the overlap between “NMD targets” and “CLIP-FMRP targets,” i.e., “NMD + CLIP-FMRP targets,” in the pool of 21,565 protein-coding genes derived from the mouse GENCODE M32 database (https://www.gencodegenes.org/) was calculated using Fisher’s exact test (P < 2.2 × 10.⁻¹⁶, odds ratio = 3.26, 95% confidence interval 2.87–3.71) in R version 3.6.3 [82]. b Box and whisker plots showing log₂ fold-changes in RNA enrichment (axons/somata) using mRNAs defined in a, after removing outliers, and data deriving from mouse retinal ganglion cells [14]. Outliers were identified, and plots were generated using the ggplot2 program (https://ggplot2.tidyverse.org/index.html) in Tidyverse package (https://www.tidyverse.org/) and R version 3.6.3. The top, middle, and bottom of each box, represent, respectively, the first, second, or third quartile of the dataset. n, number of mRNAs identified in a that existed in the dataset derived by Jung et al. [14] and was used for the statistical analysis. P values were calculated using the two-sided Wilcoxon rank-sum test using R version 3.6.3 for the full data-set, including outliers. c Cumulative fraction of log₂ fold-change RNA enrichment (axons/somata) using mRNAs defined in a, and data deriving from mouse retinal ganglion cells [14]. d As in b, box and whisker plots showing log₂ fold-changes in RNA enrichment (dendrites/somata) using mRNAs defined in a, after removing outliers, and data deriving from mouse hippocampal CA1 pyramidal neurons [13]. P values were calculated as in b. e As in c, but using data deriving from mouse hippocampal CA1 pyramidal neurons [13]