Skip to main content
Figure 4 | BMC Genomics

Figure 4

From: Deep sequencing reveals a novel class of bidirectional promoters associated with neuronal genes

Figure 4

Properties of genes associated with novel upstream-antisense lncRNAs. (a), (b), (c), (d) Distribution of Pearson correlation coefficients calculated based on the expression of protein-coding genes and nearest novel lncRNAs (red curve). The gray curves show the average correlation coefficient distribution based on 200 permutations of neighboring novel lncRNAs and protein-coding gene relationships. The gray shaded areas show standard error of the curve estimates. The distributions are shown for four possible genomic configurations of lncRNA relative to protein-protein-coding genes: upstream-sense, downstream sense, upstream-antisense and downstream-antisense. (e) GO terms enriched in 273 protein-coding genes associated with upstream-antisense novel lncRNAs. The node color indicates the GO term’s enrichment p-value, the node size is proportional to the GO term’s annotated gene number. Dashed rectangle indicates GO terms associated with neuronal functions. (f) Expression specificity of 273 protein-coding genes associated with upstream-antisense novel lncRNAs, calculated based on H3K4me3 modification from neurons and non-neural cells of human PFC. The red bar shows neuron/non-neuron cells ratio for the 273 genes, the gray bars represent the ratio distribution calculated by 1,000 permutations of 273 randomly selected expressed genes. (g) Expression specificity of 273 protein-coding genes associated with upstream-antisense novel lncRNAs calculated based on cell type specific expression data from mouse neocortex. The bars show Fisher’s test odds ratio with 95% confidence interval for enrichment of the 273 genes among mouse orthologs preferentially expressed in oligodentrocytes (blue), astrocytes (green) and neurons (red). The numbers show Fisher’s test p-values. (h), (i) The TFBS density profiles of two enriched TFs, ETF (panel h) and ZF5 (panel i), within NBiPs. The red and blue curves show observed TFBS density distributions at sense and antisense strands. The pink and the light blue curves represent TFBS density distributions calculated by 1,000 permutations of TFBS prediction across dinucleotide shuffled NBiP sequences.

Back to article page