This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Differential gene expression

Hi all, I am just a bit concerned about my data and trying to understand what that means, could I please get insights from experienced fellows? and thanks in advance! In a differential gene expression analysis can it be possible that the number of up regulated genes is zero while there are over 200 genes down regulated in a diseased tissue? I have checked my code and model multiple times it looks like all is good. Please let me know what you suggest? all the best!

deseq2 differential gene expression
• 2,645 views • link
updated by OmnibusX • OmnibusX • written by 4732b159 • 4732b159

Sure can, but one would need some plots, such as plotMA to have an idea.

• 0 views • link
written by ATpoint ATpoint

thanks for your response :) could you elaborate how to interpret that then?

• 0 views • link
written by 4732b159 • 4732b159

you could add them first here

• 0 views • link
written by ATpoint ATpoint

So,[ plotMA(res, ylim=c(-2,2)] it looks like that enter image description here

• 0 views • link
written by 4732b159 • 4732b159

1 answer

From my experience, this usually happens when I compare data from different studies, technology, or alignment tools. In these cases, the number of genes in the 2 groups is not equal, for example, in group 1, the technology can capture only 10k genes, while in group 2 the technology can capture 20k genes, during differential expression analysis, genes that are present in group 2 but not in group 1 may appear significantly downregulated. This could potentially skew your results toward downregulation. If you could provide more details about your samples, it would help in diagnosing the origin of this issue more precisely.

• 0 views • link
written by OmnibusX • OmnibusX

Hi Thanks for your reply :), really appreciated! The data that I am analyzing is from the same experiment of bulk RNA and I have two samples; disease and control. I am particularly interested in what the difference are between the same part of the tissue of disease vs control. I hope is it clearer now. Would you expect in this case? Thanks!

• 0 views • link
written by 4732b159 • 4732b159

In case you only have two samples, which method did you use for DEG analysis?

• 0 views • link
written by OmnibusX • OmnibusX

I used DESq2 and designed it as below:

count_data <- as.matrix(count_data)

dds _t<- DESeqDataSetFromMatrix(countData = round(count_data), colData = col_data, design= ~group)

dds <- DESeq(dds_t)

res <- results(dds, contrast=c("group","Disease","Control"))

extra info: The group column of col_data is a factor and consist of Disease or Control.

{{{all(colnames(count_data) %in% rownames(col_data) ==> TRUE }}}

Thanks so much!

• 0 views • link
written by 4732b159 • 4732b159

Sorry that I might not help in this case, as for the statistical test, I think we need more samples in each group (at least three) for a significant result. Therefore, I can not conclude anything yet.

• 0 views • link
written by OmnibusX • OmnibusX

Yes I do have more than three samples in each group: disease and control samples, more than 14 samples actually!

• 0 views • link
written by 4732b159 • 4732b159

I saw your MA plot and code. Everything looks fine. This outcome might be because the disease state is characterized by a general suppression of gene activity or specific pathways, rather than activating different pathways.

• 0 views • link
written by OmnibusX • OmnibusX

Log in to answer this question.

More posts like this