Why is the inclusion of RUVs not reflected in the EDA plots produced from normalised counts extracted from deseq2?
1
0
Entering edit mode
BioinfGuru ▴ 70
@yagalbi-11519
Last seen 7 days ago
Ireland

Hi all,

This may seem long winded, but I am basically asking if I should expect to be able to plot the effect of adding RUV on the counts from:

raw <- counts(dds, normalized = FALSE)
deseq <- counts(dds, normalized = TRUE) # only this one is shown in the images below
vst <- assay(vst(dds, blind=FALSE))
rlog <- assay(rlog(dds, blind=FALSE))

it seems they don't "see" the covariate columns of colData(dds) - I feel like I've badly misunderstood something

I think I may be wrong in my expectation here. I thought that if I include RUV output in a new deseqdataset then the EDA plots (pca/rle/heatmaps etc.) would reflect the inclusion of RUV. But it doesn't. I'm pretty sure I'm running the code correctly. I've also used a very high k value to make the changes obvious (although it looks like the data needs it).

So when does the effect of adding the sources of variation show up? Is it just in the results?

Here is the code:

# Before RUVs:
dds <- DESeqDataSetFromMatrix(countData = raw,
                              colData = metadata,
                              design = ~trial + trial:condition)
dds <- dds[rowSums(counts(dds) >= 10) >= 5,]
dds <- DESeq(dds)
all_plots(counts(dds, normalized = TRUE), metadata, "trial_condition", -0.5, 0.5, "Before")

# RUVs
set <- newSeqExpressionSet(counts = raw, phenoData = metadata)
differences <- makeGroups(metadata$trial_condition)
set_s <- RUVs(x = set, cIdx = rownames(set), k = 18, differences)
all_plots(normCounts(set_s), metadata, "trial_condition", -0.5, 0.5, "RUVs")

# After RUVs:
dds <- DESeqDataSetFromMatrix(countData = counts(set_s),
                              colData = pData(set_s),
                              design = ~trial + trial:condition + W_1 + W_2 + W_3 + W_4 + W_5 + W_6 +
                                W_7 + W_8 + W_9 + W_10 + W_11 + W_12 + W_13 + W_14 + W_15 +
                                W_16 + W_17 + W_18
                              )
dds <- dds[rowSums(counts(dds) >= 10) >= 5,]
dds <- DESeq(dds)
all_plots(counts(dds, normalized = TRUE), metadata, "trial_condition", -0.5, 0.5, "After")

Before RUV: enter image description here

RUV: enter image description here

After RUV: enter image description here

RUVSeq DESeq2 dese • 387 views
ADD COMMENT
2
0
Entering edit mode

Thank you! I didn't see that.

ADD REPLY

Login before adding your answer.

Traffic: 412 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6