<i>Nipped-B</i> mutant wing discs show broadly reduced gene expression by RNA-seq analysis.

<p>(A) Plot of expression level of all active genes in matched wild-type controls (P57B / Oregon R) versus <i>Nipped-B</i>^<i>407</i> / Oregon R 3^rd instar wing discs. RNA-seq was performed using ribosomal RNA depletion and normalization to histone RNA to compensate for the broad effects of Nipped-B on gene transcription [<a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1005655#pgen.1005655.ref024" target="_blank">24</a>]. Active genes (7,595) were defined as those averaging more than 500 nucleotides of sequence coverage between replicates. Pearson correlation coefficients between replicate samples for all genes quantified (14,296) ranged from 0.96 to 1.00, indicating high reproducibility (<a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1005655#pgen.1005655.s006" target="_blank">S1 Table</a>). The green line indicates a slope of 1, revealing that the <i>Nipped-B</i> mutants show diminished expression of many genes. Red circles indicate genes in which the change in RNA levels have a p value ≤ 0.05, which represent roughly a third of active genes as shown by the pie chart. (B) Density distribution plots of the fold-changes in gene expression between <i>Nipped-B</i> mutants and the wild-type control (log2 <i>Nipped-B</i>^<i>407</i> / P57B ratio) for all genes (black) and those genes for which the change has a p value ≤ 0.05. The number of genes in each group is indicated in parentheses. This shows that relative to histone RNA levels, most genes show lower RNA levels in the <i>Nipped-B</i> mutant wing discs, and that most statistically significant changes are decreases in expression. (C) Examples of expression changes for selected genes. The blue bars are set to 1 in the P57B / Oregon R control after normalization to histone RNA to allow all genes to be shown on the same scale. The error bars are the standard errors calculated from the four replicates, and p values are given above each gene. The <i>Rad21</i> (<i>vtd</i>) cohesin subunit gene and <i>Nipped-B</i> both show reductions. Several genes in the Notch signaling pathway show reductions, including the <i>Delta</i> (<i>Dl</i>) ligand, <i>Notch</i> (<i>N</i>), and direct Notch targets, including <i>vestigial</i> (<i>vg</i>) and genes in the <i>Enhancer of split</i> complex (<i>HLHm3</i>, <i>HLHm7</i>, <i>HLHmbeta</i>, <i>HLHmdelta</i>). Several genes that regulate growth also show reductions including <i>dm</i> (<i>myc</i>), <i>Tor</i>, the insulin receptor (<i>InR</i>) and the ecdysone hormone receptor (<i>EcR</i>). The <i>string</i> (<i>stg</i>) cdc25 cell cycle control gene does not show a reduction, in contrast to the <i>cyclin E</i> (<i>CycE</i>) gene that drives S phase and the <i>cyclin B</i> (<i>CycB</i>) G2 cyclin gene. The <i>Wrinkled</i> (<i>W</i>, <i>hid</i>, <i>head involution defective</i>) pro-apoptosis gene shows a significant decrease. The <i>RpL32</i> and <i>RpL7</i> ribosomal protein genes, which are often used as normalization controls for qRT-PCR experiments and Northern blots, do not show significant changes in expression. The expression values for all genes in the RNA-seq analysis are in <a href="http://www.plosgenetics.org/article/info:doi/10.1371/journal.pgen.1005655#pgen.1005655.s006" target="_blank">S1 Table</a>, along with the ontology analysis indicating that pathways involved in development, morphogenesis and growth are broadly affected, although the only significant phenotype in adult <i>Nipped-B</i> mutant wings is the reduced cell number and size. (D) qRT-PCR validation of the effects on <i>Rad21</i> (<i>vtd</i>) and <i>Nipped-B</i> RNA levels in 3^rd instar wing discs, normalizing to <i>RpL32</i> RNA levels. Ten P57B / Oregon R and eight <i>Nipped-B</i>^<i>407</i> / Oregon R replicates were assayed. The error bars are standard error calculated from all PCR replicates, and the p values for the comparisons to the controls are indicated.