Entering edit mode
3.3 years ago
Omar Mohamed
•
0
Hi, I am learning DGE. I am using DEseq2. I have obtained data from recount2. This plot comparison between 2 different cancerous tissues.
What I was expected that the dispersion is a low mean and convergence as the mean increases. (that's the correct normalization as I have learned so far)
Here is my code!
rse_genE <- read_counts(rse_gene)
covMat <- round(assays(rse_genE)$counts, 0)
#Perform DE analysis by xml_diagnosis
dds <- DESeqDataSetFromMatrix(countData = covMat, colData = colData(rse_genE), design = ~xml_diagnosis)
dds <- DESeq(dds)
plotDispEsts(dds)
res <- results(dds)
plotMA(res, main = "DESeq2 results from TCGA-lung")
Here is the image of my MA plot
What two samples are you comparing?
Both of them are cancerous cells. They are expected to be very similar. So, MA plot should be not that disperse and literally in the reverse direction across the x-axis, right?
What exact samples are they? Depending on cancer, tissue, and patient information they could be very different. It's hard to comment without any sample information though.
It depends on cancer tissue, yes. But both of them are lung cancer, but different tissue type
I figured out something. Only the grey dots are statistically significant. In that case that makes sense for these dots. However, how all the genes are that are very far in both average and fold change are not statistically significant?!