Question

DESeq2 plotCount question

0

Entering edit mode

Antonio Ahn ▴ 10

@antonio-ahn-10629

Last seen 2.8 years ago

University of Otago

Hi everyone

I am trying to do the plotCounts function in DESeq2 and have a question on how to plot for a specific gene that i am interested in.

Wherever i look the codes that is recommended is:

plotCounts(dds, gene=which.min(res$padj), intgroup="dex")

plotCounts(dds, gene="ENSG____", intgroup="dex")

What code would you use to make the plotCounts using the gene symbol instead of looking at the gene with the lowest padj or the ENSG number.

I already have a "symbol" column with the gene names.

Currently my dds or results table (res) looks something like this:

baseMean log2FoldChange lfcSE stat symbol entrez
<numeric> <numeric> <numeric> <numeric> <character> <character>
ENSG00000000003 708.60217 -0.37424998 0.09873107 -3.7906000 TSPAN6 7105
ENSG00000000005 0.00000 NA NA NA TNMD 64102
ENSG00000000419 520.29790 0.20215551 0.10929899 1.8495642 DPM1 8813
ENSG00000000457 237.16304 0.03624826 0.13684258 0.2648902 SCYL3 57147
ENSG00000000460 57.93263 -0.08523371 0.24654402 -0.3457140 C1orf112 55732

deseq2 • 2.3k views

ADD COMMENT • link updated 8.6 years ago by Michael Love 43k • written 8.6 years ago by Antonio Ahn ▴ 10

score 0 · Answer 1 · 2016-05-04

0

Entering edit mode

Michael Love 43k

@mikelove

Last seen 1 day ago

United States

What happens when you try the ENSG? See the help page ?plotCounts

ADD COMMENT • link 8.6 years ago Michael Love 43k

0

Entering edit mode

Oh I see now... you want to look up with Symbol instead.

You can provide a character to 'gene', but it has to be the rownames of the DESeqResults object.

You might then think to try setting rownames of the results table to the Symbol, but you will likely encounter an error because the Symbol often contains missing values and repeated values (ENSG is not a 1-1 mapping with Symbol), neither of which are allowed for rownames of a DataFrame.

Why not look up the ENSG id's for the genes you are interested in as a separate step, then provide these to plotCounts?

ADD REPLY • link 8.6 years ago Michael Love 43k

0

Entering edit mode

Hi Michael. Would you be able to tell me how i can get the ENSG id for the gene of interest in res$symbol ? Sorry i am quite new to R.

ADD REPLY • link 8.6 years ago Antonio Ahn ▴ 10

0

Entering edit mode

You can use the square bracket and a logical index to get the rowname for the row with a certain Symbol.

I'd recommend you take a short R course (some of these are ~1 hour, some longer). Here are some examples at the top:

http://genomicsclass.github.io/book/pages/resources.html

You can learn a lot and this will help you with basic bioinformatic tasks.

Here's an example of using square bracket to find a particular rowname:

> res <- data.frame(symbol=letters[1:4], row.names=letters[5:8])
> res
  symbol
e      a
f      b
g      c
h      d

> rownames(res)[res$symbol == "b"]
[1] "f"

For doing this many times, you could use the match() function, but this example above should get you started.

ADD REPLY • link 8.6 years ago Michael Love 43k

0

Entering edit mode

Thank you so much for your help! and also for the recommendation of the short R courses.

ADD REPLY • link 8.6 years ago Antonio Ahn ▴ 10