Computation of neutral Z when simulating varying Va values

[YannDussert]

Hi,

I would like to use the Qpc approach on some data and I had a question about the simulations incorporating environmental variation with varying Va values (Simulations-traitqpc).

In the neutQpc function, breeding values (Z) for neutral traits are simulating using:
myZneutral = mvrnorm(n=1, rep(0,dim(myK)[1]), myVamyK + myVediag(dim(myK)[1]))

However, the equation 19 (appendix 1) from the paper indicates:
vector Z ~ MVN(µ, 2VaK + Ve*I)

Could there be a missing factor of 2 in the R command?
In relation to that, it is indicated that the variance of Cm should be centred around Va, but following equation 6 (Cm ~ N(0, 2Va) ), should it be around 2Va?

I am not a skilled quantitative geneticist (and an even worst mathematician), so I apologize in advance if I missed something.

Best regards,
Yann

[emjosephs]

Hi Yann -- thanks for pointing this out. This is indeed a factor of 2 error in the code. I've updated the code in Simulations-traitqpc and plotted a new figure (while leaving the original one to compare). Let me know if you find any other issues or have any questions.
-- Em

[YannDussert]

Hi,

Thank you for your quick reply. No other question from me (yet ...).

Thanks for putting up the code and results from the paper here by the way, this is very useful.

Best regards,
Yann