For example, the splicing factor Fox-2, encoded by A-1210477 price Rbfox2, is known to be sufficient for exon inclusion in Ewsr1 in a dose-dependent manner ( Underwood et al., 2005), producing a brain-enriched isoform ( Melot et al., 2001). Consistent with this, we found the FPKM of Rbfox2 across samples was strongly correlated with both the isoform fraction (r = 0.85; p < 0.01, one-tailed) and the FPKM of the brain-enriched Ewsr1 isoform (ENSMUST00000102930; r = 0.87; p < 0.01, one-tailed). In contrast, the

total FPKM of Ewsr1 was significantly correlated neither with the isoform fraction (r = −0.47; not significant) nor with the FPKM of this brain-enriched isoform (r = 0.19; not significant). selleck kinase inhibitor Genes that exhibited patterned expression were, in general, very different from those expressed more evenly across the cortical layers. They were more likely to encode specific receptor types and proteins involved in synaptic transmission and ion transport (Figure 4A; Belgard et al., 2011). By contrast, unpatterned genes tended to possess housekeeping cellular roles that less often contribute to cell type-specific functions. An exception to this housekeeping generalization is mitochondrial genes, which were significantly and strongly enriched

in layer 5 (Table S6). We note that the extremely large layer 5 pyramidal neurons, which are the only cell types that extend axons beyond the skull, are known to have high energetic demands (Guillemot et al., 2006, Molnár and Cheung, 2006 and Molyneaux et al., 2007). Unpatterned genes appear to be more important in early development: they were significantly more likely to result in prenatal lethality when disrupted in mice (Figure 4A). Cells

in cortical layers are proposed to have differing biological roles. We sought transcriptomic evidence for these roles by identifying, for each layer in turn, molecular annotations of genes that were more abundant than expected from a bias-corrected random sampling of all 11,410 classifiable genes (Supplemental Experimental Procedures; Belgard et al., 2011). Molecular annotations were drawn from a variety of sources, including the Gene Ontology (Ashburner et al., 2000), why genomic intervals associated with human diseases and traits identified by genome-wide association studies (Chen et al., 2010), and mouse knockouts (Blake et al., 2011), and only results retained after application of a 5% false discovery rate threshold for each set are reported. Expression of genes encoding subunits of the NMDA receptor were enriched in layers 2/3: expression of all five classifiable genes were identified as being significantly concentrated in these layers, a number much higher than expected simply by chance (Figure 4C).