This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Log fold change filter using limma

This is probably a stupid question, but I felt the need to ask it as I couldn't seem to find the answer in the limma documentation. But if you wanted to filter some microarray data with a log fold change of 1.5% , what would you enter into the limma topTable function. Does the lfc parameter take percentages directly i.e should I enter lfc=1.5 or should I enter lfc=0.015.

limma log fold change toptable
• 3,556 views • link
updated by Biostar Biostar • written by flute28 • flute28

I'm not quite sure what you mean by 1.5% with respect to log-fold changes in microarray analysis. Do you want the genes that display log fold-changes in the top 1.5% of all genes studied? or do you want genes where the fitted parameter for log-fold-change is at least 1.5% of the constant coefficient in your model

• 0 views • link
written by russhh russhh

Log fold change is generally not presented with a %, where did you find that?

• 0 views • link
written by WouterDeCoster WouterDeCoster

This is not presented in term of percentage, simply fold change is a ratio between two conditions, and its normalised by log2.

for example:

logFC =   log2 (expression value of treatment / expression value of control)
     eg.  expression value of treatment = 16.2512
             expression value of control = 1.202, 
                = 16.2512/1.202 = 13.5201; 
                =log2(13.5201)
                so, logFC = 3.75

its nicely explained by Devon Ryan’s post:

R: How to convert log2FC (Fold Change) obtained by limma's topTable() function to FC

see below link also

http://rstudio-pubs-static.s3.amazonaws.com/13988_bb11d85b79b2436280de434988558140.html

• 1 views • link
written by Mike Mike

It is a minor point Mike, but you have put the ratio at the contrary. isn't ?. In addition, if you express more after a treatment, the logFC must be positive

• 0 views • link
written by Antonio R. Franco Antonio R. Franco

Yes, antonio, in that case it will be positive.

• 1 views • link
written by Mike Mike

Mike. Revise your formula. You have put the terms at the contrary. It should be 16.2512/ 1.202 , so the logFC is positive. It is a minor point , though. Sorry to disturb you

• 1 views • link
written by Antonio R. Franco Antonio R. Franco

Thanks Antonio, for catching that point..

• 1 views • link
written by Mike Mike

0 answers

No answers yet.

Log in to answer this question.

More posts like this