Data can be found here: /work/DAT_002__TCGA_Ovarian/01_2011/vita_work/CoexprCometh

Here we are investigating whether DNA methylation is the cause of coexpression modules. Solution: find correlated pairs of coexpression and comethylation PCs, find if the genes (CpGs close to genes) overlap between two correlated PCs. # of permutations: 1000

Calculation of overlap:

mb:

P value	#Raw P values	mean (permuted P val)	FDR, %
0.001	771	4.28	0.5
0.005	1050	21.1	2
0.01	1213	42.1	3.5
0.05	1765	210.5	11.9

mbc:

P value	#Raw P values	mean (permuted P val)	FDR, %
0.001	771	5.28	.68
0.005	1050	26.2	2.49
0.01	1213	52.2	4.3
0.05	1765	259.1	14.68

Is it normal that I have the same number of significant pairs for both mb and mbc for every P value cutoff?

Thanks to Justin for his advice, I also set up permutations for rho values (for a total of 1000):

mb:

Absolute rho	# of raw rho	mean (permuted)	FDR, %
0.4	49	0
0.5	26	0
0.6	18	0
0.7	8	0

mbc:

Absolute rho	# of raw rho	mean (permuted)	FDR, %
0.4	54	0
0.5	40	0
0.6	33	0
0.7	17	0

Browser not supported