Question: Longitudinal data comparing a continuous variable in limma
0
gravatar for akknight
3.2 years ago by
akknight0
akknight0 wrote:

Hi,

I'm new to limma and am trying to figure out how to test for differential expression using a continuous variable. My experimental set-up is two visits for each person and I want to look at change in expression between visit 1 and visit 2. However, I don't want to use visit 1 and visit 2 as the outcome as the visits were taken a various lengths of time apart, so I'm interested in modeling weeks post exposure (visit 1) verses weeks post exposure (visit 2). ( I want to see if there's an association with weeks) I've tried:

group<-pheno$Subject
cont.var<-pheno$Weeks
X<-splines::ns(cont.var,df=5)
design<-model.matrix(~0+group+X)
fit <- lmFit(edata2, design)
fit2<-eBayes(fit)
a<-topTable(fit2,coef="group", n=Inf, adjust="fdr")

(But this gives me output for each subject, not weeks)

Or

group<-pheno$Visit
cont.var<-pheno$Weeks
X<-splines::ns(cont.var,df=5)
design<-model.matrix(~group+X)
fit <- lmFit(edata2, design)
fit2<-eBayes(fit)
a<-topTable(fit2,coef="X5", n=Inf, adjust="fdr")

(But this seems equivalent to testing visit 1 verses visit 2) 

Unfortunately, I can't post the data here.

edata2 is a dataframe of log transformed expression values (Illumina platform)

Any ideas? 

 

ADD COMMENTlink written 3.2 years ago by akknight0
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 2.3.0
Traffic: 1556 users visited in the last hour