Validated pipeline

Bulk RNA-seq differential expression

Differential gene expression between conditions from bulk RNA-seq data.

Input
Paired-end RNA-seq reads from samples in two or more conditions.
Output
The log2 fold change and its significance for every gene, for each comparison between conditions.

Validation

We analysed the publicly available RNA-seq data of the reference study, which compared dexamethasone-treated with untreated primary human airway smooth muscle cells from four donors, and compared each of the 28 results reported in that study with the corresponding result of this reproduction. All 28 results were reproduced (Figure 1). The log2 fold changes of the 17 genes in Table 1 of the reference study agreed closely with the reported values (Pearson r = 0.9988; Figure 2). The number of differentially expressed genes (q < 0.05) was 300, compared with 316 reported, and 278 of the 316 reported genes (88.0%) were differentially expressed in the same direction in this reproduction (Figure 3). The mapping rate (%) of every sample was within or above the reported range (Figure 4).

Figure 1. Reproduction of results reported in the reference study (n = 28)

Reproduction of results reported in the reference study (n = 28) 28 symbols, one per reported result: all reproduced. Number of differentially expressed genes (q < 0.05): reproduced. 300 versus 316 reported C7 log2 fold change: reproduced CCDC69 log2 fold change: reproduced DUSP1 log2 fold change: reproduced FKBP5 log2 fold change: reproduced GPX3 log2 fold change: reproduced KLF15 log2 fold change: reproduced MAOA log2 fold change: reproduced SAMHD1 log2 fold change: reproduced SERPINA3 log2 fold change: reproduced SPARCL1 log2 fold change: reproduced C13orf15 log2 fold change: reproduced TSC22D3 log2 fold change: reproduced CRISPLD2 log2 fold change: reproduced PER1 log2 fold change: reproduced KCTD12 log2 fold change: reproduced ERRFI1 log2 fold change: reproduced STEAP4 log2 fold change: reproduced Overlap with the reported differentially expressed genes: reproduced CEBPD: differentially expressed, increased: reproduced CEBPB: expressed, not differentially expressed: reproduced SYNPO2: differentially expressed: reproduced B2M: highly expressed, not differentially expressed: reproduced GABARAP: highly expressed, not differentially expressed: reproduced GAPDH: highly expressed, not differentially expressed: reproduced RPL19: highly expressed, not differentially expressed: reproduced IL1B: low expression in both conditions: reproduced Mean mapping rate (%): reproduced 28 of 28 results reproduced
Show data table
Reported resultOutcomeValues
Number of differentially expressed genes (q < 0.05)Reproduced300 versus 316 reported
C7 log2 fold changeReproduced
CCDC69 log2 fold changeReproduced
DUSP1 log2 fold changeReproduced
FKBP5 log2 fold changeReproduced
GPX3 log2 fold changeReproduced
KLF15 log2 fold changeReproduced
MAOA log2 fold changeReproduced
SAMHD1 log2 fold changeReproduced
SERPINA3 log2 fold changeReproduced
SPARCL1 log2 fold changeReproduced
C13orf15 log2 fold changeReproduced
TSC22D3 log2 fold changeReproduced
CRISPLD2 log2 fold changeReproduced
PER1 log2 fold changeReproduced
KCTD12 log2 fold changeReproduced
ERRFI1 log2 fold changeReproduced
STEAP4 log2 fold changeReproduced
Overlap with the reported differentially expressed genesReproduced
CEBPD: differentially expressed, increasedReproduced
CEBPB: expressed, not differentially expressedReproduced
SYNPO2: differentially expressedReproduced
B2M: highly expressed, not differentially expressedReproduced
GABARAP: highly expressed, not differentially expressedReproduced
GAPDH: highly expressed, not differentially expressedReproduced
RPL19: highly expressed, not differentially expressedReproduced
IL1B: low expression in both conditionsReproduced
Mean mapping rate (%)Reproduced
Each symbol is one result reported in the reference study: the log2 fold changes of the 17 genes in Table 1, the overlap with the reported list of differentially expressed genes, eight statements on the expression of individual genes, the mean mapping rate and the number of differentially expressed genes. Filled circles denote results reproduced in this analysis. All 28 results were reproduced. The number of differentially expressed genes (q < 0.05) was 300, compared with 316 reported.

Figure 2. Concordance of log2 fold changes for the 17 genes in Table 1 of the reference study

Concordance of log2 fold changes for the 17 genes in Table 1 of the reference study Scatter plot of 17 genes: reported log2 fold change on the horizontal axis, reproduced log2 fold change on the vertical axis, with the identity line. Pearson r = 0.9988. -3 -2 -2 -1 -1 0 0 1 1 2 2 3 3 4 4 5 5 6 6 Reported (log2 fold change) Reproduced (log2 fold change) y = x C7: 3.35 reported, 3.21 reproduced CCDC69: 2.92 reported, 2.87 reproduced DUSP1: 2.99 reported, 2.95 reproduced FKBP5: 3.95 reported, 3.92 reproduced GPX3: 3.76 reported, 3.72 reproduced KLF15: 4.58 reported, 4.57 reproduced MAOA: 3.29 reported, 3.3 reproduced SAMHD1: 3.83 reported, 3.84 reproduced SERPINA3: 3.34 reported, 3.28 reproduced SPARCL1: 4.7 reported, 4.6 reproduced C13orf15: 3.27 reported, 3.17 reproduced TSC22D3: 3.27 reported, 3.24 reproduced CRISPLD2: 2.7 reported, 2.69 reproduced PER1: 3.2 reported, 3.19 reproduced KCTD12: −2.52 reported, −2.49 reproduced ERRFI1: 2.43 reported, 2.44 reproduced STEAP4: 4.99 reported, 5.22 reproduced FKBP5 CRISPLD2 STEAP4 KCTD12
Show data table
GeneReportedReproducedDifference
C73.353.21−0.14
CCDC692.922.87−0.05
DUSP12.992.95−0.04
FKBP53.953.92−0.03
GPX33.763.72−0.04
KLF154.584.57−0.01
MAOA3.293.30.01
SAMHD13.833.840.01
SERPINA33.343.28−0.06
SPARCL14.74.6−0.1
C13orf153.273.17−0.1
TSC22D33.273.24−0.03
CRISPLD22.72.69−0.01
PER13.23.19−0.01
KCTD12−2.52−2.490.03
ERRFI12.432.440.01
STEAP44.995.220.23
Each point is one gene; the horizontal axis gives the log2 fold change reported in the reference study and the vertical axis the log2 fold change in this reproduction (dexamethasone-treated versus untreated; n = 4 donors). The diagonal marks identity (y = x). All 17 estimates fall within 0.24 log2 units of the reported values (mean absolute difference 0.054; Pearson r = 0.9988). The largest difference is for STEAP4 (5.225 versus 4.99 reported).

Figure 3. Reproduction of the 316 genes reported as differentially expressed (q < 0.05)

Reproduction of the 316 genes reported as differentially expressed (q < 0.05) A bar of the 316 reported differentially expressed genes, divided into 278 reproduced with the same direction of change and 38 not reproduced. 316 genes reported as differentially expressed (q < 0.05) 278 reproduced (same direction) 38 not reproduced 278 reproduced (same direction) 38 not reproduced
Show data table
CategoryGenesPercentage
reproduced (same direction)27888.0%
not reproduced3812.0%
Total316100%
The bar represents the 316 genes reported as differentially expressed (q < 0.05) in the reference study (Table S3), divided according to whether each gene is also differentially expressed in this reproduction with the same direction of change. Of the 316 genes, 278 (88.0%) were reproduced and 38 (12.0%) were not.

Figure 4. Mapping rate per sample relative to the reported range

Mapping rate per sample relative to the reported range 8 points, one per sample, on a mapping-rate axis (%), over a band for the reported range of 81.9% to 84.3% and a line at the reported mean of 83.4%. Reported range: 81.9% to 84.3% Reported range 81% 82% 83% 84% 85% 86% Reported mean: 83.4% Reported mean, 83.4% SRR1039508: 85.9% SRR1039509: 84.6% SRR1039512: 85.9% SRR1039513: 85.5% SRR1039516: 84% SRR1039517: 85.2% SRR1039520: 85.5% SRR1039521: 85.8% This reproduction (n = 8)
Show data table
SampleValue (%)
SRR103950885.9%
SRR103950984.6%
SRR103951285.9%
SRR103951385.5%
SRR103951684%
SRR103951785.2%
SRR103952085.5%
SRR103952185.8%
Reported range81.9% to 84.3%
Reported mean83.4%
Each point is one sample (n = 8; four donors, treated and untreated). The shaded band marks the per-sample range reported in the reference study (81.9–84.3%) and the vertical line the reported mean (83.4%). The mean mapping rate in this reproduction is 85.3% (range 84.0–85.9%). Of the 8 samples, 7 lie above the reported range, 1 within it and 0 below it.

Reference

Himes BE, Jiang X, Wagner P, et al. RNA-Seq transcriptome profiling identifies CRISPLD2 as a glucocorticoid responsive gene that modulates cytokine function in airway smooth muscle cells. PLoS One. 2014;9(6):e99625. doi:10.1371/journal.pone.0099625

Back to all pipelines