I saw that HyPrColoc assumes independence between the studies by default (it is much faster and performed well in simulations compared to the correlated model).
Changing it between TRUE and FALSE in the test example does not affect the results. However, adding trait.cor to the default example did change the result (and run much slower).
Hi James,
I saw that HyPrColoc assumes independence between the studies by default (it is much faster and performed well in simulations compared to the correlated model).
When looking into the hyprcoloc function source code, I found hyprcoloc <- function(... ind.traits = FALSE)
Changing it between TRUE and FALSE in the test example does not affect the results. However, adding trait.cor to the default example did change the result (and run much slower).
On the other hand, when looking into the sensitivity.plot function source code, ind.traits are set to TRUE by default.
I was wondering if ind.traits should be also set to TRUE by default in hyprcoloc function?
Best wishes, Oleg