This page describes correlation of DNA methylation data in ovarian cancer with clinical traits. The most critical trait here is age since it has been published that there are 589 loci are correlated with age in ovarian cancer (however, tho whole blood was used to do the profiling). The loci were selected using Spearman correlation (for quantitative traits) and Kruskal-Wallis test for categorical traits. Threshold is 5%.
trait |
Level 2 |
Level 1 - lumi |
snm (meth. probes only) |
|
|---|---|---|---|---|
stage |
137 |
6 |
|
NA |
grade |
127 |
0 |
|
NA |
residual disease |
0 |
6 |
|
NA |
therapy outcome success |
0 |
0 |
|
NA |
age |
2423 |
2200 |
|
589 |
days to death |
14435 |
0 |
|
NA |
days to tumor recurrence |
0 |
0 |
|
NA |
We are trying to evaluate which method is the most appropriate for normalization. Here we are going to compare each method "performance" by identifying the number of loci associated with age and finding an overlap and associated P-values with each overlap. The top right triangle represents the overlap and the bottom left triangle represents the significance of overlap (P-values)
|
Level 2 |
Level 1 - lumi |
snm (meth.probes only) |
Teschendorff, 2010 |
|---|---|---|---|---|
Level 2 |
|
1689 |
NA |
101 |
Level 1 - lumi |
0 |
|
NA |
89 |
snm (meth.probes only) |
NA |
NA |
|
NA |
Teschendorff, 2010 |
2.261381e-11 |
1.763328e-09 |
NA |
|
Overlap between level 1- lumi/Teschendorff, 2010 and level 2/Teschendorff, 2010 is 75.