Dear people,

I am trying to discover whether there is a correlation between my continuous variable X, and the change in multiple metabolites in response to a diet (three different diets were followed).

I started by looking at Pearson correlation which did give some results. However, I realized that the diet itself had already shown to be a predictor for the change in variable X and should thus be corrected for.

So in other words: I need to correct the Pearson correlation for the change in variable X that was caused by the diet, what is the best way to do this? Should I use partial correlation, multiple regression analysis, co variate analysis or something else?

ps. There are over a hundred metabolites to be tested, so a procedure that could test multiple dependent variables at once would be helpful.