This LaTeX document is available as postscript or asAdobe PDF.

ANSC*6370 QG and Animal Models
Fall 03 - Quiz 10 - Answers

1.

Non-Additive Genetic Model

(a)

How might you simulate dominance genetic effects?

You need a complete dominance genetic covariance matrix for all animals that you wish to simulate, say ${\bf D}$. Then do a Cholesky decomposition of ${\bf D}$ into ${\bf L}'{\bf L}$. Generate a vector of random normal deviates of length equal to the number of animals, call it ${\bf w}$, then the vector of dominance genetic effects would be

$${\bf d} = {\bf L}'{\bf w}.$$

(b)

What assumption is necessary for the non-additive genetic model that was discussed in class?

That there has been no selection or culling of animals - the population is randomly mating.

2.

Maternal Effects Model

(a)

If $\sigma^{2}_{a}=10$, $\sigma^{2}_{m}=4$, $\sigma_{am}=-2$, $\sigma^{2}_{mpe}=25$, and $\sigma^{2}_{e}=100$, then calculate the appropriate k-values for use in the mixed model equations for the maternal genetics effect model.

$$\left( \begin{array}{ll} k_{a} & k_{am} \\ k_{am} & k_{m} \en... ...\end{array} \right), \ \mbox{and} \ k_{pe}=\frac{100}{25} = 4.$$

(b)

Do sires receive an estimated breeding value for maternal genetic effects? The answer is Yes - but tell me how this happens.

Through the relationship matrix and their daughters that have progeny and express the maternal genetic merit, and through ancestors evaluated for maternal effects.

This LaTeX document is available as postscript or asAdobe PDF.