Entering edit mode
7 months ago
Pumla
•
0
Hello Biostar community,
I've been encountering an unusual output in my volcano plots, and I'm seeking some insights from the community. Typically, all my graphs tend to come out facing downwards. I've used the following codes for analysis:
# Change resLFC to a dataframe
resLFC <- as.data.frame(resLFC)
# Label the genes
resLFC$diffexpressed <- "NO"
resLFC$diffexpressed[resLFC$log2FoldChange > 0.1 & resLFC$padj < 0.05] <- "UP"
resLFC$diffexpressed[resLFC$log2FoldChange < 0.1 & resLFC$padj < 0.05] <- "DOWN"
resLFC$delabel <- NA
ggplot(data = resLFC, aes(x = log2FoldChange, y = log10(pvalue), col = diffexpressed, label = delabel)) +
geom_point() +
theme_minimal() +
geom_text_repel() +
scale_color_manual(values = c('blue', 'black', 'red')) +
theme(text = element_text(size = 20))
I would appreciate any insights or suggestions on why my volcano plots are consistently facing downwards and any potential adjustments or corrections to the code.
Thank you in advance for your assistance and expertise.
It's
-log10(pvalue)
, you miss the minus. Some people use indeed pvalue, but since you're filtering on padj, I would show-log10(padj)
rather than pvalue. It's better to interpret in terms of common cutoffs.what are common cutoffs?
Perhaps more concerning is why you have features (genes?) with small pvalue but fold-change very close to zero. It's also unusual that there is a gap between features with ~0 lfc and features with lfc ~ |2|. Not saying it's wrong, just unusual.