Glucocorticoid gene expression data

JL Jonathan Lu
BD Bianca Dumitrascu
IM Ian C. McDowell
BJ Brian Jo
AB Alejandro Barrera
LH Linda K. Hong
SL Sarah M. Leichter
TR Timothy E. Reddy
BE Barbara E. Engelhardt
request Request a Protocol
ask Ask a question
Favorite

We analyzed RNA-sequencing data from a set of experiments developed to study glucocorticoid receptors (GRs) in the human adenocarcinoma and lung model cell line, A549 [6]. There was an original exposure data set of 4 replicates in which cells were stimulated by the glucocorticoid dexamethasone (dex), and gene expression was profiled at {0, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12} hours of dex stimulation. There was also an unperturbed data set of 3 replicates in which cells were exposed to dex for 12 hours, after which the conditioned media was replaced and dex removed. Gene expression was profiled at {0, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12} hours after dex removal. We integrated the original exposure and unperturbed data into a joint data set with 7 replicates. The original exposure data set is available at the Gene Expression Omnibus (GEO), with reference numbers listed for the rows that list “RNASeq” as the Assay under the column “Experiment_GEO_Series” in Supplementary Table 3 of [6]. The GEO accession numbers for time points {0, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12} of the original exposure data set are GSE91305, GSE91198, GSE91311, GSE91358, GSE91303, GSE91243, GSE91281, GSE91229, GSE91255, GSE91284, GSE91222, and GSE91212, respectively. The unperturbed data set is available at GEO accession number GSE144662.

We selected 2768 genes for analysis, which had average expression > 2 Transcripts Per kilobase Million (TPM) and were differentially expressed in the original exposure data. A gene was called differentially expressed if its expression at any time point differed from its expression at time 0, ascertained by running edgeR (FDR ≤ 0.05) [6]. We added NR3C1, which encodes the glucocorticoid receptor (GR). NR3C1 was not found to be differentially expressed at FDR ≤ 0.05.

After genes were selected, gene expression TPM were log-normalized and corrected for surrogate variables using SVAseq [85]. Each gene’s temporal profile was centered to have mean zero across time. In the original exposure data, all replicates besides replicate 1 had a measurement for each time point. Replicate 1 was missing time points 5 and 6 hrs, so we imputed these values using a linear interpolation from time points 4 and 7 hrs in the log-transformed, surrogate-corrected space.

Do you have any questions about this protocol?

Post your question to gather feedback from the community. We will also invite the authors of this article to respond.

post Post a Question
0 Q&A