Search
Question: DESeq2 rna seq analysis
0
gravatar for gv
2.3 years ago by
gv30
United States
gv30 wrote:

Hi,

I am trying to use DESeq2 package. I have some question's related to that.

1. DESeq2 manual say's in MAplot function  Points will be colored red if the adjusted p value is less than 0.1. Can i change the value to 0.05. If yes how??

2. Also point which falls out of the window are plotted as open triangles. In the file what should I see in the column of adjusted p value. Will it be NA??

 

Thanks

Varun

 

ADD COMMENTlink modified 2.3 years ago by Steve Lianoglou12k • written 2.3 years ago by gv30
1
gravatar for Steve Lianoglou
2.3 years ago by
Genentech
Steve Lianoglou12k wrote:

1. DESeq2 manual say's in MAplot function  Points will be colored red if the adjusted p value is less than 0.1. Can i change the value to 0.05. If yes how??

In the help page you get from ?plotMA function for a DESeqDataSet, you'll see that there is an alpha parameter you can tweak which sets the significance threshold to plot.

2. Also point which falls out of the window are plotted as open triangles. In the file what should I see in the column of adjusted p value. Will it be NA??

The position on the y-axis is defined by the log fold change of the gene expression, what you'll see in the related adjusted p value column is not a function of the fold change, but rather its p-value, so there isn't a real "direct" answer to your question (although on average you'd expect high logFC's with low p-values, but it's not always the case), if you know what I mean?

You should note that NA pvalues come from two sources (if you've run DESeq2 w/o any tweaks):

  1. From genes that have outlier expression in >= 1 sample; or
  2. From genes that have been filtered out by the "independent filtering" mojo DESeq does by default to reduce the multipe-testing burden via removing genes that have a small chance of being differentially expressed
ADD COMMENTlink modified 2.3 years ago • written 2.3 years ago by Steve Lianoglou12k

So in the file, how can i distinguish red points with red triangles which are outliers?? THis is because I need to know which genes are really differentially expressed and is not an outlier

ADD REPLYlink modified 2.3 years ago • written 2.3 years ago by gv30

Which file?

The object you get from calling results has columns for the log2 fold change as well as the adjusted pvalue -- those two columns together will etll you which ones are "really" differentially expressed, no? If there are NA's in the adjusted p values, these may or may not be outliers -- you will have to go back to your data an inspect it, the plotCounts function will be helpful here.

ADD REPLYlink written 2.3 years ago by Steve Lianoglou12k

Ok I will look at the plotCounts function, it's usage and get back to you. Yes I was referring to the object generated after applying results function

res<-results(dds)
res<-res[order(res$padj),]
ADD REPLYlink written 2.3 years ago by gv30

@Steve: So I looked at plotCounts function and plotted it for some genes. What should I look for in them so as to distinguish outliers from other differentially expressed gene.

ADD REPLYlink written 2.3 years ago by gv30

I will throw this question back to you and ask you what you mean by "outlier" here.

ADD REPLYlink written 2.3 years ago by Steve Lianoglou12k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.2.0
Traffic: 169 users visited in the last hour