Agriculture Reference
In-Depth Information
1.0
B
Gompertz
0.5
BW
70 d
mean PD
ADG
Lean content
0.0
FI
100 kg
Duration
Backfat thickness
Age at slaughter
ADFI
FCR
FI
50 kg
-0.5
-1.0
-1.0
-0.5
0.0
0.5
1.0
Component
1
(34.23%)
Fig. 2.5.
Correlation among model parameters that describe feed intake and growth in InraPorc. The main
model parameters are indicated by solid lines and derived traits by dashed lines. (From Vautier, 2013.)
feed intake curve. The shape parameter of
the Gompertz function for protein depos-
ition was positioned perpendicular to 'Pm',
indicating that there are different trajector-
ies by which pigs can attain a final protein
mass. Information about how fast an animal
grows therefore provides little information
about how it grows. The covariance struc-
ture of the
40
subpopulations was affected
by cross-breed, sex and batch (Vautier, 2013),
and meant that not only the absolute value
of model parameters will vary for each
population, but also the covariance struc-
ture. This was a rather surprising observa-
tion because it meant that each situation is
unique. It has been our experience that
users already have difficulties obtaining
data for the dynamics of feed intake and
growth for the average pig in the herd, but
quantifying the variation within a herd
would probably be beyond reach for most
users of InraPorc. Vautier (2013) determined
a median covariance matrix as the next-best
(and feasible) solution to account for vari-
ation among individuals. This median co-
variance matrix can be used to generate a
virtual population of pigs while respecting
the covariance structure of the parameters.
The algorithms used in InraPorc and the ever-
increasing computing speeds make it pos-
sible to run a simulation for thousands of
pigs within a few seconds. Vautier (2013)
evaluated whether the prediction of the
variation in a simulated population of 2000
pigs (using the median covariance matrix)
corresponded to that actually observed in a
real population. The
in vivo
and
in silico
pigs were offered feed
ad libitum
or at a re-
stricted level with a maximum of 2.4 (gilts)
or 2.7 (barrows) kg/day. Two experiments
were conducted in which the pigs were
either group-housed (by sex) or housed indi-
vidually. In both experiments, the coeffi-
cient of variation was similar for the
in vivo
