Seeking Help with Unexpected MA Plot Output

Question

[Pumla]

Hey Biostar Community,

I've been working on analyzing some RNA-seq data, and I've run into an unexpected issue with my MA plot. It's not coming out as I anticipated, and I'm having trouble figuring out what might be causing the problem.

The MA plot I generated seems off, and I'm not sure why. I've double-checked my code and data processing steps, but I can't seem to pinpoint where things might have gone wrong. The plot doesn't reflect the typical pattern I'd expect to see, and I'm concerned that there might be an error in my analysis pipeline.

I was wondering if any of you have encountered similar issues with MA plots before or if you have any insights into what might be causing this unexpected output. Any advice, suggestions, or troubleshooting tips would be greatly appreciated.

this is the code i used:

#MA plot
plotMA(dds,ylim=c(-2,2))

Thanks in advance for your help!

[jared.andrews07]

What do you feel is unexpected or "looks wrong" here? This looks pretty typical for a comparison with low power and/or few DE genes to me. Limiting the y-axis range to such a degree is the weirdest thing about it to me.

[Istvan Albert]

To me, it looks like the plot has unexpected data horizontal lines that indicate some sort of systematic bias

there are unexpectedly many points with the exact same fold changes yet varying averages at fold changes around 2 and -2

those look odd and suspect - though at least some bands seem to have different symbol, triangle instead of cross, so it might be plotting something else

[Dave Carlson]

I believe the horizontal lines at LFC 2 and -2 are artifacts of the y axis limits. Per the DESeq2 vignette:

In DESeq2, the function plotMA shows the log2 fold changes attributable to a given variable over the mean of normalized counts for all the samples in the DESeqDataSet. Points will be colored blue if the adjusted p value is less than 0.1. Points which fall out of the window are plotted as open triangles pointing either up or down.

Presumably those data points represented by triangles mostly would not fall on a line if the y limits were larger.

[swbarnes2]

Right, but there are more little lines. But since they are only visible when the normalized mean is < 10 they are probably artifacts caused by really low counts

[Istvan Albert]

This would be a good answer, actually, if the OP were worried about the top and bottom lines,

Applying a limit to a plot changes the symbols and rescales them into a different location may be unexpected.

It is a useful feature, no question about that, but it can be confusing.

[LauferVA]

please tell us everything you can about the experimental design. were data pooled? are there replicates of certain kinds? what was tested? how many samples?