Question

Vertical line at A=0, nothing else, on MA plot

0

Entering edit mode

david.wheeler7 ▴ 20

@davidwheeler7-19959

Last seen 5.8 years ago

Hello!

I am trying to generate an MA plot with the following code:

dds <- DESeq(dds)

res = results(dds, contrast=c("group","Control","653"))

resultsNames(dds)# this returns 'Intercept' 'group_653_vs_111' 'group_Control_vs_111'

plotMA(res, ylim = c(-5, 5))

This returns a plot with one vertical line that stretches from the max of the y-limit to -2 @ x=0.

Has anyone seen this before? If so, do you have any insight to resolve the issue? I imagine it is a problem with my data, not with DESeq2 per se.

Thank you in advance for your help!

David

deseq2 • 1.4k views

ADD COMMENT • link updated 5.8 years ago by Michael Love 43k • written 5.8 years ago by david.wheeler7 ▴ 20

score 0 · Answer 1 · 2019-03-14

0

Entering edit mode

Michael Love 43k

@mikelove

Last seen 19 hours ago

United States

First, I think you probably want to swap the order of the second and third arguments, if you want 653 / Control. See ?results

Can you print res, like so:

res

And copy here what the output is?

ADD COMMENT • link 5.8 years ago Michael Love 43k

0

Entering edit mode

Thanks Michael,

Sure. here is the output from res after changing the order to results(dds, contrast=c("group","653","Control"))

   log2 fold change (MLE): group 653 vs Control 
Wald test p-value: group 653 vs Control 
DataFrame with 246300 rows and 6 columns
                             baseMean     log2FoldChange             lfcSE
                            <numeric>          <numeric>         <numeric>
Cluster-67248.142691 3.47666008870489  0.366755604078382   1.8741909745588
Cluster-67248.107952 132.652518797792  0.197213624532245 0.301284552039947
Cluster-58782.0      0.53966481500208  -1.64945749337118  4.08052381408439
Cluster-67248.17374  173.196102707328  0.170057399573535 0.206834433434017
Cluster-67248.56631   4.3499365353681 -0.847111227266121  1.71641222115821
...                               ...                ...               ...
Cluster-69843.0       1.2590688096371  -1.30830606148772  2.12287329941037
Cluster-67248.66210  1.90112609039418 -0.927086840942579  1.50700584689153
Cluster-67248.129845 1.25892938295144   2.22519657083048  2.43144036615263
Cluster-67248.27096  20.5225603579964 -0.434990333494745 0.464039759789789
Cluster-67248.132887  41.631160475336  0.879490130174741 0.449698978108395
                                   stat            pvalue              padj
                              <numeric>         <numeric>         <numeric>
Cluster-67248.142691    0.1956874240976 0.844854828223016                NA
Cluster-67248.107952  0.654575958830097 0.512740806957619 0.962323811856868
Cluster-58782.0      -0.404226900398888 0.686045874056607                NA
Cluster-67248.17374   0.822190951236295 0.410968227939485 0.939615751763184
Cluster-67248.56631  -0.493536003078855 0.621633907203962                NA
...                                 ...               ...               ...
Cluster-69843.0      -0.616290224127416 0.537702989743021                NA
Cluster-67248.66210   -0.61518463438935 0.538432794101421                NA
Cluster-67248.129845  0.915176288839649 0.360099090285026                NA
Cluster-67248.27096  -0.937398842055682 0.348553436363289 0.917445133632772
Cluster-67248.132887    1.9557307732257 0.050496877612509 0.565298039900576

Do you see anything fishy?

This gave me a very similar MA plot

ADD REPLY • link 5.8 years ago david.wheeler7 ▴ 20

1

Entering edit mode

Can you use:

DESeq2::plotMA(res)

It may be that another package is masking our function.

ADD REPLY • link 5.8 years ago Michael Love 43k

0

Entering edit mode

Yes!!!! That did it!!! I do not understand the source of the error. If you have time, could please direct me toward the resources need to explain this? Thanks again Michael!

ADD REPLY • link 5.8 years ago david.wheeler7 ▴ 20

0

Entering edit mode

Some package masked plotMA. R will tell you in the console when this happens, you just have to pay attention to the loading messages.

ADD REPLY • link 5.8 years ago Michael Love 43k

0

Entering edit mode

Thanks Michael! I was running it in Jupyter Notebook and did not see anything. I will try again from the console.

ADD REPLY • link 5.8 years ago david.wheeler7 ▴ 20