Log2Fold change interpretation - Continuous predictor variable
Entering edit mode
Last seen 21 days ago
United Kingdom


I was hoping someone may be able to help me with interpretation of Log2Fold change and its specific interpretation when used with a continuous predictor variable.

I understand that the Log2Fold change when used with a predictor variable should be interpreted as change per unit of the variable. However, my question is regarding the use of thresholds. When using with a continuous variable I am finding that my results are significant (adjusted p-value <0.05) log2fold changes are relatively small (<1) and I wondered whether this is likely due to the fact that I am using a continuous predictor variable as opposed to a categorical variable. I.e., would we find that a continuous predictor variable would have smaller log2fold change compared with a categorical predictor variable given the way in which it is interpreted? I am not too sure if this at all makes sense, I am a relative novice when it comes to this sort of analysis. So it may just be that the log2 fold changes are small and the differences are negligible across the continuous variable, but I just wanted to seek some advice.

Any help would be greatly appreciated.

Thank you!

DESeq2 • 118 views
Entering edit mode
Last seen 1 day ago
United States

It can be more sensitive than binary versions, this has been written about in statistical literature.

But most of all I would encourage use of lfcThreshold in results(). We’ve been encouraging this argument since we posted DESeq2 paper in 2013, for all designs actually.

Entering edit mode

Thank you very much for your help.


Login before adding your answer.

Traffic: 341 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6