RadMeth Output Meaning

[Bonnie Cantrell]

Hello,

The output for Radmeth includes a column called "meth-diff". Is this value a percentage? If it is a percentage, would a sample with -0.498148, as stated in the manual, have a 49.8% or 0.49% difference in methylation? 

Also, some of my results have a methylation difference that is extremely close to zero. The number of results like this decreased when I change the FDR from 0.05 to 0.01, but I still see some that are basically 0. How can there be a significant difference when the difference is close to zero? 

Thank you,

Bonnie Cantrell 

[Ben Decato]

Hi Bonnie,

In our modeling, the methylation level ranges from 0 to 1. A methylation difference of -0.498148, then, corresponds to a -49.8% difference.

As the number of samples in your study increases, the power to detect significant changes at smaller effect sizes increases. If you have a reasonable sample size and are testing few enough regions, it's not unreasonable to see significant p-values at small effect sizes. I would caution against over-interpreting these small effect sizes though: generally I don't think methylation differences less than 10% are meaningful, or would be unwilling to try and correlate them to transcriptional changes (for example) without single-cell resolution. Usually my first follow-up to a radmeth run is to plot a volcano plot (effect size vs -log10 p-value) and decide on a threshold for effect size.

If you think your experiment is small enough that the p-values you're seeing are erroneous, it would be helpful to see the command you used to run, the design matrix, and a snippet of the proportion table for debugging purposes.

Thanks,

Ben

--
You received this message because you are subscribed to the Google Groups "MethPipe and MethBase Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to methpipe+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

-- 
Benjamin Decato, Ph.D.
Computational Biology Section,
Department of Biological Sciences,
University of Southern California
1050 Childs Way, RRI 408M
Los Angeles, CA 90089-2910