Computational protocol: Reversible Block of Mouse Neural Stem Cell Differentiation in the Absence of Dicer and MicroRNAs

Similar protocols

Protocol publication

[…] MicroRNA levels in Dicer-null NS cells were normalised to those in wild-type NS cells using the set of control small RNAs on the array (tRNAs, snRNAs). MicroRNA levels were expressed relative to wild-type levels, correcting for local and systemic background signals. Gene expression differences between wild-type and Dicer-null self-renewing NS cells were identified by comparing two wild type and two Dicer-null NSC lines on six dye-swapped MEEBO microarrays: four hybridisations directly compared expression between Dicer-null and wild-type lines, and two hybridisations directly compared the two wild-type lines. Expression analysis was carried out using the Limma package. Low intensity features were identified and weighted appropriately, arrays were lowess normalised and then normalised between arrays. Control features and features not representing genes were excluded before the identification of differentially expressed genes using the eBayes algorithm. An adjusted p-value cut-off of 0.05 was used to identify significant genes. Gene ontology analysis of the gene expression changes in Dicer-null NS cells was performed using GOToolBox ( . GO categories (p<0.01) significantly enriched by comparison with the whole probeset were identified using a Benjamini & Hochberg corrected hypergeometric test.To compare gene expression changes between wild-type and Dicer-null NS cells during glial differentiation, RNA was harvested from wild-type and Dicer-null NS lines 24 hours after the addition of BMP4 or PBS. A set of six array hybridisations was carried out: two hybridisations comparing gene expression between wild-type NS cells with and without BMP4 treatment; two hybridisations comparing gene expression between Dicer-null NS cells with and without BMP4 treatment; and two hybridisations directly comparing BMP-treated wild-type and Dicer-null NS cells. Array data were lowess normalised (in Acuity) and the entire set of six arrays analysed as a group. Data were filtered to remove low-intensity features and features absent on more than two arrays. To identify robust changes in gene expression occurring in at least one of the cell lines, only genes showing at least a 50% change in gene expression in two or more arrays were retained for further analysis. The content of this final set of genes was explored by hierarchical and agglomerative clustering in the Acuity system to identify the major classes of gene expression changes, as reported above. To prospectively identify differences in the response to BMP4 treatment between wild-type and Dicer-null NS cells, we used the significance analysis of microarrays (SAM) algorithm to analyse the four microarrays comparing BMP-treated and control Dicer-null and wild-type NS cells. Data were normalised and filtered as before and significant changes in gene expression discovered by performing a two-class SAM, using a false discovery rate (FDR) cut-off of 0.1. Data passing the 0.1 FDR were hierarchically clustered using MEV to aid visualisation . […]

Pipeline specifications

Software tools limma, GOToolBox, SAM
Application Gene expression microarray analysis
Organisms Mus musculus