Question: should I get the logarithm of count values when I use standard method
0
gravatar for Bioinformatics
16 months ago by
Bioinformatics30 wrote:

I have a question which is very important for me. 

Basically I am buidling my condition like this 

condition <- factor(c(rep("Normal", 10),rep("Tumor", 11)))​

The values are count values and I want to know if I am doing the condition correct and if so, should I take the logarithm 2 myself before I use the standard method ?

I mean by standard method is dds <- DESeq(dds)

dseq2 • 232 views
ADD COMMENTlink modified 16 months ago by Simon Anders3.6k • written 16 months ago by Bioinformatics30
Answer: should I get the logarithm of count values when I use standard method
0
gravatar for Simon Anders
16 months ago by
Simon Anders3.6k
Zentrum für Molekularbiologie, Universität Heidelberg
Simon Anders3.6k wrote:

No, of course not! Why would you think that?

(Background: The DESeq manual states, as explicitly as possible and since the very first version, that the matrix should contain "read counts", i.e., integer numbers that count how many reads map to the gene. Nevertheless, I have been asked hundreds of time various variations of: I have RPKM values. Can I use these, too? My data has been normalized and now it's no longer integer. Can I still use hem? I have done [insert arbitrary mathematics operation] to my counts. I can still input this to DESeq, right?  etc.

So, please use DESeq on read counts and only on read counts. And take a moment to recall the meaning of the word "count". ;-)

Sorry for the rant; you just triggered with your question my sort-of-annual reminder to this site about this issue.)

 

As for your condition: Do you have paired samples, i.e., on tumour and one normal sample per patient? If so, you need to let DESeq know which samples belong to the same patient. Read up on "paired designs" in the manual.

ADD COMMENTlink modified 16 months ago • written 16 months ago by Simon Anders3.6k

@Simon Anders I did not mean to offend you. The problem is that if I use count , then I don't know if it will be normalized insdie the standard calculation or not. I read the vinegette very well but to be honest, I have never seen such an explanation, I even read their paper . That is why I ask here. 

No they are not pair. so is my condition correct when I want to compare Treated versus Control ?

Thanks a lot for your comment

ADD REPLYlink written 16 months ago by Bioinformatics30
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 16.09
Traffic: 122 users visited in the last hour