Output different from Galaxy than when run in R

usegalaxy.org support
1

I ran DESeq2 on usegalaxy.org and generated the figure on the left. I then took the VST file exported on Galaxy and ran the following code on R to generate a similar heatmap:

sampleDists <- dist(t(assay(vsd)))
library("RColorBrewer")
sampleDistMatrix <- as.matrix(sampleDists)
rownames(sampleDistMatrix) <- paste(vsd$condition, vsd$type, sep="-")
colnames(sampleDistMatrix) <- NULL
colors <- colorRampPalette( rev(brewer.pal(9, "Blues")) )(255)
pheatmap(sampleDistMatrix,
         clustering_distance_rows=sampleDists,
         clustering_distance_cols=sampleDists,
         col=colors)

As you can see from the figures, the distance on the clustering looks the same for both, but the groupings are drawn differently. Which is correct? Any idea what could have happened?

heatmaps
heatmaps1920×1080 109 KB

2

Hi @mitchellgc

The Bioconductor forum has much discussion about the different normalization methods. Start with these DESeq2 normalization vs VST vs rlog && Clarification re VST transformed data in DESeq2 && choosing the normalization method (rlog, variance stabilizing transformation).

3

Deseq 2 nomalization method is based on deseq2 median ratio. You are using different normalization method in R

1 Like