coded such that higher scores imply higher trait levels). Yet, if the true genotype-phenotype model is a network model, contrasting GV effects are unproblematic from a statistical point of view. Whether contrasting effects are plausible from a biological perspective depends on the actual functional role of the GV. For instance, symptoms like blunted affect and agitated mood can both be positive indicators of depression on a population level, but their biochemical basis may be antagonistic, making contrasting GV effects for these symptoms both statistically and biologically possible.
