I performed analysis of association of single DNA methylation loci with the survival. Justin was very helpful in constructing correct survival object (censoring information included days to the last follow-up and days to death). There were very few loci associated with survival. Therefore we were interested if the entire comethylation modules would be predictive of survival. I didn't make a Sweave file with step by step analyses because it was very similar to the single locus analysis (the same survival object). In this case I used first principal components of each module. Used Wald's p-values. Here are the plots of the p-values obtained in Cox proportional hazards model for each module. Left image is for mb, right image is for mbc.

It seems that we have some but weak association. Permutation analysis for the mb dataset:

P value	Targets	Permuted, mean	FDR,%
0.01	1	0.62	62
0.05	6	3.19	53
0.005	0
0.001	0

Permutation analysis for the mbc dataset:

P value	Targets	Permuted, mean	FDR,%
0.01	2	0.77	38.5
0.05	8	3.95	49.4
0.005	0
0.001	0

Browser not supported