Question: iClusterPlus no significant results
0
gravatar for Pin.Bioinf
21 months ago by
Pin.Bioinf270
Malaga
Pin.Bioinf270 wrote:

Hello,

Is anyone familiar with iClusterPlus?

I have been trying to use iClusterPlus to integrate methylation and expression data I have, but It's not clear to me what I might be doing wrong:

I am using beta values and RPKMS.

1) I filter out genes with low rpkm variability and probes (cpgs) with low beta value variability using kOverA function. 2) I use the tuning function to select the better results like:

set.seed(123)

   for(k in 1:5){
    cv.fit = tune.iClusterPlus(cpus=2,dt1=as.matrix(methx),dt2=as.matrix(exprx),
                                type=c("gaussian","poisson"),K=k,n.lambda=35,
                                 scale.lambda=c(1,1,1),maxiter=20)  
   save(cv.fit, file=paste("cv.fit.k",k,".Rdata",sep=""))
  }

3) My % of explained variation plot depending on the number of clusters is an increasing plot, (the more number of clusters, the most % explained variation), so maybe I should try more clusters than only 5, so I could see a better classification.

4) The clusters generated are shown in a heatmap: I can see the CpGs variation but for expression, the heatmap is almost everything the same color. Might this mean that I am filtering way too many genes?

I am using 17798 genes and 244 cpgs (maybe I need more cpgs)

How can I select the genes and probes belonging to each cluster? I can't seem to find the way by looking at the manual, everytime I get the same genes and probes.

Thank you so much.

ADD COMMENTlink written 21 months ago by Pin.Bioinf270
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: 1609 users visited in the last hour