Computational protocol: Neuroblastoma cells depend on HDAC11 for mitotic cell cycle progression and survival

Similar protocols

Protocol publication

[…] Microarray data obtained in duplicates were normalized using quantile normalization in the 'R programming language'. Quantile-normalized Illumina mRNA data were log2 transformed. Differentially expressed transcripts were identified for each HDAC11 siRNA separately in comparison to the negative control transfection using the empirical Bayes approach as implemented in the Bioconductor package limma. Time points were tested globally using moderated F-statistics and individually using moderated t-statistics, both based on the same linear model. All P-values were adjusted for multiple testing using the Benjamini–Hochberg correction. Next, transcripts that showed a significant regulation at two or more time points in the same direction and no significant regulation in the opposite direction, were selected. The overlap of these selected transcripts from both siRNAs defined the primary list of regulated transcripts (hits). This analysis was performed separately for each cell line. GO terms were analyzed for over-representation and enrichment. First, hypergeometric tests were used to test for over-representation of GO terms within the hit list. Second, a gene set enrichment analysis was performed using the minimal F-statistics from both siRNA models as a global measurement of regulation. Taking the minimal moderated F-statistic value from both linear models can be considered a conservative approach because the lesser strength of regulation from both siRNAs is selected to represent the transcript, thus mimicking the requirement to show regulation in both siRNAs. In case multiple transcripts mapped to the same Entrez Gene ID, duplicates were removed for both analyses by using only the transcript that showed the strongest regulation to represent the gene. Again, P-values were adjusted for multiple testing using Benjamini–Hochberg correction. GO analyses were carried out for each cell line separately with the Bioconductor package HTSanalyzeR. GO terms showing both a significant over-representation and enrichment were selected for further investigation. All analyses were carried out using R, and all tests were two-sided.Effects of HDAC11 depletion on phenotype compared to those of negative control siRNA transfection were analyzed by a mixed linear model with a fixed effect for HDAC11-depleted samples in comparison to negative control treated cultures and random intercept for each individual HDAC11 or negative control siRNA using SAS PROC MIXED, SAS Version 9.2 (SAS Institute Inc., Cary, NC, USA). Comparison of qRT-PCR data was performed with a paired two-tailed t-test (GraphPad Prism version 5.01, GraphPad Software, Inc., La Jolla, CA, USA). P-values below 0.05 were considered statistically significant. […]

Pipeline specifications

Software tools limma, HTSanalyzeR
Databases Gene
Application Non-coding RNA analysis
Organisms Homo sapiens