Agriculture Reference
In-Depth Information
and because this method can work with
limited number of observations. The double
exponential forecasting equation, which
produces an
i
i-period-ahead forecast at time
t
was evaluated (Hauschild
et al
., 2012) for
pigs using individual DFI and WT data col-
lected from nine performance tests (2406
pigs) at the CDPQ Experimental Station in
Deschambault (Quebec, Canada) and more
recently calibrated (Rivest
et al
., 2012).
The mechanistic component of the
model was used to estimate starting day nu-
trient requirements and optimal dietary
amino acid concentrations for each pig in the
herd based on starting day expected DFI (or
NE intake), BW and DG information previ-
ously estimated by the empirical model com-
ponent. Daily protein deposition (PD) is as-
sumed to be a given proportion of DG (PD/
DG); this proportion evolving over time ac-
cording to a function calculated from trials
that were held at Lennoxville's research farm
(Rivest
et al
., 2013, unpublished results). In
this mechanistic model component, BW, DFI
and DG are the driving variables used in a
factorial procedure estimating the amino
acid requirements (g/day) of each individual
animal using the relationships described by
Cloutier
et al
. (2013) and previously pre-
sented in this chapter. The optimal concen-
tration of these amino acids (g/NE) that
should be provided in its daily individually
tailored diet is then calculated by dividing the
sum of the maintenance and growth require-
ments by the expected NE intake. At this point
in time, other nutrient requirements, including
minerals or vitamins, are not yet explicitly
estimated in the described model.
Thus, the empirical model component
uses each pig's up-to-date data to estimate the
starting day expected DFI, BW and DG val-
ues, while these forecasted values are then
used by the mechanistic model component
to estimate the standardized ileal digestible
lysine (SID Lys) and other amino acid re-
quirements, as well as the optimal concentra-
tion of these nutrients in the feed for the starting
day. An initial evaluation of the model (Haus-
child
et al
., 2012) was completed with data from
a study that investigated the effect of feeding
pigs with a three-phase or daily-multiphase
system. The mechanistic model component
estimated the optimal SID Lys/NE ratio with
reasonable between-animal (average coeffi-
cient of variation (CV) = 7%) and over-time
(average CV = 14%) variation. Thus, the
amino acid requirements estimated by the
model are animal- and time-dependent and
follow, in real time, the individual DFI and
BW growth patterns. It was concluded that
the proposed model can follow the DFI and
BW trajectories of each individual pig in real
time with good accuracy (Hauschild
et al
.,
2012). Based on these trajectories and using
classical factorial equations, the model esti-
mates dynamically the amino acid require-
ments of each individual animal. The factor-
ial method was calibrated in two animal trials
(Zhang
et al
., 2011, 2012; Cloutier
et al
., 2013)
and the overall approach of estimating real-
time amino acid requirements was challenged
in two validation trials (Andretta
et al
., 2014).
This model has recently been updated (Rivest
et al
., 2012; Rivest
et al
., 2013, unpublished
data) and will soon be evaluated in commer-
cial conditions.
First calibration trial
The mechanistic component of the mathem-
atical model estimating in real time the in-
dividual Lys requirement was calibrated
with growing (from
25
to 55 kg of BW) and
finishing (from
70
to 100 kg of BW) pigs. For
this calibration trial, four experimental diets
were mixed daily in different proportions to
provide each pig with a diet containing
60,
70,
80,
90,
100
or 110% of the estimated Lys
requirements while maintaining adequate
levels of all other nutrients. The four diets
were formulated on the basis of NE and
apparent ileal digestible (AID) amino acids.
The two sets of diets, A
1
/A
2
and B
1
/B
2
, only
differed in the level of added crystalline Lys
and were all formulated to have a minimum
of
10
MJ of NE per kilogramme of feed.
Feed A
1
was formulated to satisfy the re-
quirements of the most demanding pigs at
the beginning of the first growing period
and B
1
those of the less demanding pigs at the
end of the last growing period (NRC, 1998).
Feeds A
2
and B
2
contained, respectively,
the same amounts of nutrients as A
1
and B
1
