I want to do a correlation analysis of proteins to a clinical variable.
e.g. correlation of a protein to weight
For many individuals I have: 4 time points with different weights 4 protein intensities.
I can now make correlation tests and will get p-values for proteins for individuals
Person A: Protein 1 p=0.1 Protein 2 p=0.01 ...
Person B: Protein 1 p=0.2 Protein 2 p=0.3 ...
...
How can I calculate, if Protein 1 correlates significantly within the whole data set with weight? I have now as many p-values as individuals for every protein?
Is this the right approach?
I think i found the answer myself.
a t-test with p-values from the correlations.
The null hypothesis: $H_0= $ pvalue $\geq 0.05$