Correction: A RNA-Seq Analysis of the Rat Supraoptic Nucleus Transcriptome: Effects of Salt Loading on Gene Expression

There are a number of errors in the captions for S1 to S9 Fig. in the Supporting Information. Please see the complete, correct S1 to S9 Fig. captions here. S1 Fig. Micrograph illustrating magnocellular neurons (MCNs) in a salt loaded SON mounted on a PENmembrane frame slide and visualized by a Arcturus XT laser capture microscope. Large blood vessel adjacent to the SON is shown in the upper left corner. Scale line is 100um. S2 Fig. RNA-Seq Analysis Workflow. (A) Overview of analysis steps performed. Step 1, "Sequence Adaptor Clipping”. Read pairs, 101bp in length, were adaptor clipped via FASTQ/A Clipper (http://hannonlab.cshl.edu). Broken read pairs as a result of clipping were discarded. Step 2, “Sequence Quality Inspection”. Intact read pairs remaining post-clipping were import into CLCbio and the “Create Sequencing QC Report” tool used to generate one report per sample. These reports, each containing per-sequence and per-base quality statistics, were individually inspected and cross-compared to define universal hard-trimming rules to be applied to all samples. Step 3, “Sequence Quality Trimming & Filtering”. The CLCbio “Trim Sequences” tool was used to hard-remove the first 15nt from the 5’ end and the last 1nt from the 3’ end of each read pair. The tool was also used to dynamically-trim away nucleotides having a call accuracy rate less than 95%. Read pairs having at least one sequence containing more than two ambiguities were also discarded as part of this Step as were read pairs having at least one sequence with a post trimmed length less than 15 nucleotides. Step 4, “Sequence Alignment and Enumeration”. The CLCbio “RNA-Seq Analysis” tool was used to align read pairs to the Rat Genome (RN5) by sample using default parameters. Output provided by the tool included a Reads per kilo base per million (RPKM) expression value for 26,313 genes. Step 5, “RPKM Expression Pedelstalling”. Output from Step 4 was imported into R (http://www.r-project.org/) and a value of two added to each RPKM expression value per sample. Step 6, “RPKM Expression Transformation”. Pedestalled values from Step 5 (RPKM+2) were Log2 transformed using standard commands in R then filtered to keep only those genes having a post-transformed expression value (Log2(RPKM+2))>1 for at least one sample. Step 7, “RPKM Expression Normalization”. Transformed values from Step 5 were quantile normalized using standard commands in R. Step 8, “Exploratory Analysis”. Normalized values from Step 7 (Quantile(Log2 (RPKM+2))) were interrogated in R by Tukey box plot, covariance-based principal component analysis (PCA) scatter plot and Pearson correlation-based heat map to confirm absence of outliers. Step 9, “Noise Analysis”. For each gene, the coefficient of variation (CV) and mean expression was calculated by sample class using standard commands in R then modeled by sample class using the lowess() command. Step 10, “Confidence Criterion Selection”. Lowess fits from Step 9 were visually inspected to define the mean expression value across sample