That makes sense.
Do you think in this case it is a good option to add the flag --illcond to run this analysis without getting an error? or is it better not to use the variable apoe in these statistical analyzes?
What do you recommend us?

Best wishes,
Marina.