Question: Surrogate variable analysis: do I need to adjust all SVs?
gravatar for wwyue127
5 months ago by
wwyue12710 wrote:

I am analyzing the HumanMethylationEPIC array data. In order to remove batch effects and other unwanted variation, surrogate variable analysis (SVA) was conducted using the SmartSVA package. And then include the surrogate variables (SVs) as covariates in the linear model.

Question: If the sample size is small (for example, only 40), and there are many SVs (for example, 15). Do I need to adjust all SVs? If not, what criteria should I use to select the number of SVs?

I'm waiting for your help! Thanks!

ADD COMMENTlink modified 12 weeks ago • written 5 months ago by wwyue12710
gravatar for wwyue127
12 weeks ago by
wwyue12710 wrote:

I found a paper (DOI: 10.1186/s13229-018-0224-6) that says,"We chose the number of SVs to include in the model by considering both the number of SVs at which the estimated lambda values began to plateau and where the known potential confounders appeared to be captured by one or more SVs."

But a professor in Harvard said he would use all SV, that is how the method is designed...

I am still strugglingI and i dont know what to do

ADD COMMENTlink written 12 weeks ago by wwyue12710
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 991 users visited in the last hour