Strange MA plot from DESeq2

Question

[Yuyayuya]

Hello guys,

I got a strange MA plot like this.

The code I used was simply like the manual:

plotMA(res)

Does anyone know the reason and how to fix it?

Thank you in advance!

[Kevin Blighe]

It looks like it's plotting virtually the same thing against itself. The log ratios should neither be that high, and one would expect both positive and negative log ratios.

How did you produce res (go through all of the DESeq2 commands)?

[Yuyayuya]

Here's the code I used:

lncap_de <- DESeq(data_deseq_1)
lncap_de_out <- results(lncap_de)
res <- results(lncap_de, alpha = 0.05)

And, I checked the result by:

head(res[ order(resSig$log2FoldChange), ])

It looks like (log2FoldChange seems normal):

I tried demo data from DESeq and got the same strange MA plot as well.

dds <- makeExampleDESeqDataSet()
dds <- DESeq(dds)
res <- results(dds)
plotMA(res)

[Kevin Blighe]

I repeated your commands for the demo data and my MA plot looked normal.

You may want to respond to Devon's comment, and provide your session's info via sessionInfo().

I run R version 3.2.5 (2016-04-14) on Platform: x86_64-pc-linux-gnu (64-bit), under Ubuntu 14.04.5 LTS, with DESeq2_1.10.1

[Yuyayuya]

Thanks for replying!

Here's the session's info: R version 3.4.0 (2017-04-21) Platform: x86_64-apple-darwin15.6.0 (64-bit) Running under: macOS Sierra 10.12.6, with DESeq2_1.16.1

Answer 1

[Yuyayuya]

For those people encounter the same problem like me.

Type plotMA in R and check the working environment. For example, I got

> plotMA
function (object, ...) 
UseMethod("plotMA")
<environment: namespace:limma>

It means the plotMA() I used isn’t DESeq2’s plotMA function. In this situation, use DESeq2::plotMA() instead of plotMA() to make MA plot.

[Kevin Blighe]

Thanks for sharing the update! That's certainly something about which to be careful in the future.

[jambler24]

This worked for me too, had the same problem

[rohitsatyam102]

This worked for me too

Answer 2

[Devon Ryan]

It looks like your model lacked an intercept and you didn't specify a contrast. You'll need to show the full set of commands (specifically, how you made your original DESeqDataset and the contents of any sample sheet involved). Also, you'll need to post the output of sessionInfo().

[Kevin Blighe]

Hey Devon, you may want to take a look above!

[Yuyayuya]

Thanks for replying!

I generated the DESeqDataset by combining countData and colData by DESeqDataSetFromMatrix( )

Here's the entire code I used:

 > dat = read.csv(file = "/User/Desktop/gene_count_matrix.csv", header = T, row.names=1)
 > colData = read.table(file = "/User/Desktop/colData.txt", header = T, row.names=NULL)
 > n_count <- rowSums(dat) 
 > data <- dat[n_count > 0, ]
 > rownames(colData) <- colnames(data)
 > data_deseq_1<- DESeqDataSetFromMatrix(countData =
            data, colData = colData,
            design = ~ Genotype + Trt + Time)
 > lncap_de <- DESeq(data_deseq_1)
 > lncap_de_out <- results(lncap_de)
 > res <- results(lncap_de, alpha = 0.05)

Here's how input .csv and .txt data looks like:

I don't understand why even the MA plot of demo data also looks strange.

[Devon Ryan]

At this point I haven't a clue what's going on. I suspect there's simply something installed incorrectly on your system. I suggest you post this on the bioconductor site and wait for Mike Love (the DESeq2 author) to reply. Perhaps he has more insight as to where the problem might be.

[Yuyayuya]

Thank you, Sir! I'll do that. It may be true since I did have some error when I installed DESeq2.