Omics explorer 2

Normalized data were filtered on flags to exclude probes, which have missing values in more than one sample, and these were tissue-wise quantile normalized using the Limma Package of R program (www.r-project.org). Normalized data were analyzed using Agilent’s Genespring GX v12.0, the Limma package of R, and Qlucore Omics Explorer 2.3 (Qlucore AB, Lund, Sweden). Each brain region at each time point (T5R1, T5R10, T10R1, and T10R42) was compared for aggressor exposure effects: Aggressor-Exposed (Agg-E) vs. Control (C-ctrl). DEGs in each tissue and time point were identified using a Moderated T-test at p < 0.05 of the Limma Package of R.
Time effects were also compared using TimeClust [64 (link)] across different combinations of aggressor exposure sessions and rest periods (T5R1, T5R10, T10R1, and T10R42).

Transcriptomic profiling was carried out on whole blood RNA as described previously [7 (link)] using Affymetrix GeneChip Human Genome U133 plus 2.0 Arrays. For background correction, the Robust Multichip Average was applied with quantile normalisation and a mean probe set summarisation using Qlucore Omics Explorer 2.3 ([QOE] Qlucore, Lund, Sweden). The data set generated was subject to quality control to investigate the presence of outliers and further confounding effects.
Baseline gene expression associations with height velocity in each year of growth response were determined using rank regression with microarray batch, age, body mass index (BMI) at baseline as covariates (eliminated factor function in QOE) for both GHD and TS patients along with gender and peak GH test response (average of two provocative tests) for the GHD patients. Over the study, a number of children either entered puberty spontaneously or received exogenous sex steroids for pubertal induction. We therefore introduced a further normalisation for Tanner stage to the analysis to account for the proportion of children entering puberty in each year of the study.

Principal component analysis (PCA) plots were generated using data where the expression level of each gene had been standardized to zero mean and unit variance by Qlucore Omics Explorer 2.3 (Qlucore).

In order to associate functional gene sets to each lncRNA in the top of the profiles, we performed gene set enrichment analysis (GSEA). We used the highest ranked lncRNA of the profiles, and computed the Pearson correlation coefficient for each lncRNA-mRNA combination. mRNAs were then ranked according to the Pearson correlation coefficient to generate ranked gene lists for GSEA. GSEA was performed by the JAVA program [30 ] using the MSigDB C2 CP: REACTOME gene set collection (674 gene sets). Gene sets with a false discovery rate (FDR) value <0.05 after performing 1,000 permutations were considered significant [31 (link)].
Relative expression levels of the top lncRNAs, in different subtypes, both in our dataset and in the Affymetrix validation datasets were visualized by box plots and tested for significant associations with the molecular subtypes using the t-test. All plots and tests were performed in Qlucore Omics Explorer 2.3 (Qlucore).

Log₂ standardized abundances were used and duplicate injections were merged. Hierarchical clustering analysis was performed using Qlucore Omics Explorer 2.3 software (Qlucore AB, Lund, Sweden). Multi-group comparison was done using analysis of variance (ANOVA) and complete linkage was used for hierarchal clustering. Supervised classification was performed with a support vector machine. Leave one out cross-validation was employed and only binary classifiers were used with no seed. Every sample was left out once, and the remaining samples were used as the training set. The sample left out was tested on the resulting classifier and a decision value was obtained. Large positive and negative decision values correspond to predictions in the two classes respectively. Varying the threshold decision value between the two classes produces a set of classifiers. Corresponding values of sensitivity and selectivity are plotted as a receiver operating characteristic (ROC) curve. The library e1071 (http://cran.r-project.org/web/packages/e1071/index.html) was used for the support vector machine. Default parameters and a linear kernel were used for the classifier. Only peptides without missing values were used for the classifier and values were logarithmically transformed and normalized. All statistics was performed in R. ROC curves are provided in Additional file 6: Figure S1.