Animal Science

http://journals.cambridge.org/ASC
Additional services for Animal

Science:

Email alerts: Click here

Subscriptions: Click here
Commercial reprints: Click here
Terms of use : Click here

The use of compositional growth curves for assessing the

response to dietary lysine by high-lean growth gilts
K. G. Friesen, J. L. Nelssen, R. D. Goodband, M. D. Tokach, A. P. Schinckel and M. Einstein
Animal Science / Volume 62 / Issue 01 / February 1996, pp 159 - 169
DOI: 10.1017/S1357729800014430, Published online: 02 September 2010

Link to this article: http://journals.cambridge.org/abstract_S1357729800014430

How to cite this article:
K. G. Friesen, J. L. Nelssen, R. D. Goodband, M. D. Tokach, A. P. Schinckel and M. Einstein (1996). The use
of compositional growth curves for assessing the response to dietary lysine by high-lean growth gilts. Animal
Science, 62, pp 159-169 doi:10.1017/S1357729800014430
Request Permissions : Click here

Downloaded from http://journals.cambridge.org/ASC, IP address: 128.210.126.199 on 30 Jul 2014

Animal Sriraiv 11%, 62: 159-169

1W6 British Society of Animal Science

1357-7298/96/45160159$20-00

The use of compositional growth curves for assessing the response to

dietary lysine by high-lean growth gilts
K. G. Friesen1, J. L. Nelssen 1 , R. D. GoodbancTt, M. D. Tokach1, A. P. Schinckel2 and M. Einstein2
'Kansas State University, Manhattan, KS, USA
Purdue University, West Lafayette, IN, USA

Abstract
Growth modelling was used to characterize the response to digestible lysine in two experiments (114 gilts in
experiment 1 and 96 gilts in experiment 2) from 34 to 72-5 kg and 72-5 to 136 kg, respectively. Maize-soya-bean
meal diets were formulated to assure that lysine (5-4 to 10-4 and 5-4 to 9-4 g digestible lysine per kg for experiments
1 and 2, respectively) was the first limiting amino acid. Analysis of variance was used to test linear and quadratic
responses in cumulative weight gain on test as digestible lysine increased. A time X digestible lysine interaction
(linear, P < 0-001) was detected, indicating that a separate regression equation for each lysine level was necessary.
In experiment 1, average daily gain (ADG) and carcass crude protein (CP) accretion were maximized for gilts
given 10-4, 9-4 and 8-4 g digestible lysine per kg from 34 to 44 kg, 44 to 54 kg, and 54 to 72-5 kg, respectively. Lipid
accretion was minimized for gilts given 7-4 to 8-4 g digestible lysine per kg. In experiment 2, ADG was maximized
by feeding 8-4 g/kgfrom 72-5 to 92-5 kg and 7-4 g/kgfrom 92-5 to 136 kg. Carcass CP accretion was maximized by
feeding 9-4 g digestible lysine per kg, whereas lipid accretion was minimized for gilts given 8-4 g digestible lysine
per kg from 72-5 to 136 kg. Iffeeding graded levels of digestible lysine resulted in parallel lines for protein accretion,
mean values would result in accurate data evaluation. However, responses to digestible lysine changed over the
feeding period. Therefore, the use of body weight and compositional growth curves offers an approach to more
accurately characterize the growing pig's response to increased digestible lysine.
Keywords: carcass composition, gilts, growth curve, lysine.

concept has attempted to increase the flexibility of

amino acid requirements for growing-finishing pigs
by setting indispensable amino acids in a ratio
relative to lysine (Wang and Fuller, 1989). This
concept potentially could be limited to changes in the
ratio with maturity, efficiencies of amino acid use,
and for maintenance v. growth (Benevenga, Gahl,
Crenshaw and Finke, 1994). Furthermore, the ideal
protein ratios are based upon nitrogen (N) retention
experiments,
which
may
proportionately
overestimate amino acid requirements by as much as
0-05 (Baker, 1986). An approach for modelling body
weight (BW) gain (Whittemore, Tullis and Emmans,
1988; Schinckel, 1992) has been developed accurately
to characterize total BW and tissue weight gain over
time. This modelling approach can conceptually
improve the estimates for nutrient requirements to
maximize lean tissue deposition. The objective of our
research was to use these mathematical techniques to
characterize the changing response to digestible
lysine in high-lean growth gilts given food from 34 to
72-5 kg and 72-5 to 136 kg.

Introduction
Currently, nutrient requirements for growingfinishing pigs are based on mean growth
performance or lean tissue deposition rates over a
given time period (Agricultural Research Council,
1981; National Research Council (NRC), 1988). These
static estimates of nutrient requirements limit the
flexibility accurately to formulate diets for the daily
changes in nutrient needs as well as genotype,
environment, and health status (Baker, 1986; Watt,
DeShazer, Ewan, Harrold, Mahan and Schwab, 1987;
Williams, Stahly and Zimmerman, 1994). Cook (1991)
indicated that average daily gain (ADG) and food
efficiency (G/F) can overestimate the methionine
requirement for growing-finishing pigs compared
with mathematical modelling techniques. Therefore,
these data suggest a need for improved techniques to
determine nutrient requirements. The ideal protein

+ To whom correspondence should be addressed.

159

160

Friesen, Nelssen, Goodband, Tokach, Schinckel and Einstein

Material and methods

Diet formulation

Animals and housing

The gilts and experimental procedures for these

experiments have been described previously by
Friesen, Nelsson, Goodband, Tokach, Unruh, Kropf
and Kerr (1995a and b). Experiment 1 was conducted
from August to November, 1992 and experiment 2
was conducted from April to August 1993. Briefly,
114 (experiment 1) or 96 (experiment 2) high-lean
growth gilts from a synthetic terminal sire line (Pig
Improvement Co., L326, Franklin, KY, USA) were
used in each of two experiments to determine the
digestible lysine requirement to maximize ADG and
carcass protein accretion from 34 to 72-5 kg
(experiment 1) and from 72-5 to 136 kg (experiment
2). The gilts in experiment 1 (30 kg BW) were
delivered to the Kansas State University Swine
Teaching and Research Center and given a 9 g total
lysine per g diet until they reached a mean weight of
34 kg. In experiment 2 the gilts were given a 11-5 g
total lysine per kg diet from 32 to 72-5 kg before the
experiment was initiated. Three gilts were housed
per pen (4-6 m X 1-2 m pens with solid flooring) in
an open-fronted building with six replicate pens per
treatment. Each pen contained a single whole feeder
and nipple waterer to provide ad libitum access to
food and water respectively.

Digestible lysine treatments were 5-4, 6-4, 7-4, 8-4, 9-4

and 10-4 g/kg (7-5, 7-8, 9-6, 10-3, 11-7, 12-8 g total
lysine per kg, respectively) for experiment 1 (Table 1)
and 5-4, 6-4, 7-4, 8-4 and 9-4 g/kg (7-3, 8-7, 10-5, 10-7
and 11-3 g total lysine per kg, respectively) for
experiment 2 (Table 2). The maize-soya-bean meal
ratio was adjusted to provide the desired digestible
lysine levels. Then, digestible tryptophan, threonine,
methionine + cystine, and isoleucine concentrations
were set using an ideal amino acid ratio (Chung and
Baker, 1992). Apparent ileal digestible amino acid
coefficients were taken as: maize = 0-64, soya-bean
meal = 0-85, and L-lysine-HCl = 0-93 (Knabe, LaRue,
Gregg, Martinez and Tanksley, 1989). L-lysine-HCl
was maintained at 0-5 g/kg complete diet. The
dietary metabolizable energy (ME) content was
increased to 14-3 MJ/kg by adding 30 g soya-bean oil
per kg. All other nutrients were formulated in excess
of NRC (1988) estimates for the 20 to 50 kg and 50 to
110 kg pig for experiments 1 and 2, respectively.
Carcass tissue accretion rates

Six gilts were selected randomly for slaughter at 34

and 72-5 kg for experiments 1 and 2, respectively, and
the right side of each eviscerated carcass was ground
to determine contents of moisture, crude protein (CP),

Table 1 Experiment 1 diet composition (as-fed basis)

Digestible lysine (g/kg)

Ingredient (g/kg)
Maize
Soya-bean meal (485 g CP per kg)
Soya-bean oil
L-lysine HCI
L-threonine
DL-methionine
L-tryptophan
Monocalcium phosphate (210 g P per kg)
Limestone
Salt
Trace mineral pre-mixt
Vitamin pre-mix}:
Chemical analyses (g/kg)
Crude protein (CP)
Total lysine
Ca
P

MJ/kg

5-4
796-2
140-8
30-0
0-5
0-01
160
9-5
3-5
1-5
2-0

1460
7-5
7-5
6-5
14-3

10-4

7-4

8-4

9-4

756-9
180-9
30-0
0-5
0-1
0-01

717-1
221-1
30-0
0-5
0-3
0-3

677-4
261-2
30-0
0-5
0-3
0-7

637-3
301-4
30-0
0-5
0-7
1-1

597-5
341-5
30-0
0-5
1-0
1-4

15-3
9-3
3-5
1-5
2-0

14-6
9-2
3-5
1-5
2-0

13-8
90
3-5
1-5
2-0

131
8-9
3-5
1-5
2-0

12-4
8-7
3-5
1-5
20

153-4
7-8
7-5
6-5
14-3

171-2
9-6
7-5
6-5
14-3

190-8
10-3
7-5
6-5
14-3

201-5
11-7
7-5
6-5
14-3

208-6
12-8
7-5
6-5
14-3

6-4

t Provided the following per kg of complete diet: Mn, 100 mg; Fe, 100 mg; Zn, 100 mg; Cu, 10 mg; I, 3 mg; Co, 1 mg; Se, 3 mg.
X Provided the following per kg of complete diet: retinol, 992 ng; cholecalciferol, 8-3 |ig; a-tocopherol; 8-7 mg; riboflavin, 3-3 mg;
D-pantothenic acid, 8-3 mg; niacin, 18-2 mg; choline, 331 g; cyanocobalamin, 0-02 mg; menadione (menadione sodium bisulphate
complex), 1-3 mg.

Values from calculated analyses on as-fed basis.

Growth modelling in pigs

161

Table 2 Experiment 2 diet composition (as-fed basis)

Digestible lysine (g/kg)

Ingredient (g/kg)
Maize
Soya-bean meal (485 g CP per kg)
Soya-bean oil
L-lysine HCI
L-threonine
Di.-methionine
Monocalcium phosphate (210 g P per kg)
Limestone
Salt
Trace mineral pre-mixt
Vitamin pre-mix:):
Chemical analyses (g/kg)
Crude protein (CP)
Total lysine
Ca
P

MJ/kg

54

6-4

788-3
148-9
30-0
0-5

749-0
1890
30-0

15-8
9-5
3-5
1-5
2-0

132-2
7-3
7-5
6-5
14-3

0-5
0-08
0-02
15-1
9-3
3-5
1-5
2-0

151-4
8-7
7-5
6-5
14-3

9.4

7-4

8-4

709-2
229-1
30-0
0-5
0-3
0-36
14-4
9-2
3-5
1-5
2-0

669-3
269-3
300
0-5
0-52
0-72
13-7
9-0
3-5
1-5
2-0

629-2
309-4
30-0

160-7
10-5
7-5
6-5
14-3

174-3
10-7
7-5
6-5
14-3

189-6
11-3
7-5

0-5
0-83
1-2
130
8-9
3-5
1-5
2-0

6-5
14-3

See Table!.
lipid, and ash (Association of Official Analytical
Chemists (AOAC), 1990). When the mean weights of
gilts in a pen were approximately 54 and 72-5 kg
(experiment 1) and 104 to 136 kg (experiment 2), one
pig from each pen (six pigs per treatment) was
slaughtered for carcass analyses. The head, leaf fat,
and viscera were removed at slaughter and were not
included in determination of tissue accretion rate. At
24 h post mortem, the right side of each carcass was
ground once through a 15-mm plate and once through
a 9-mm plate and homogenized for 3 min in a ribbonpaddle mixer. Proximate analyses (AOAC, 1990) were
conducted in triplicate on each carcass sample. From
the chemical analyses, the amounts (g/kg) of CP,
lipid, ash and dry matter (DM) were determined for
each carcass based upon cold carcass weight. Moisture
content was determined by subtracting the DM (g/kg)
from 1000. Thus, initial composition, determined from
chemical composition of carcass weight, was
subtracted from chemical composition determined at
55 and 72-5 kg (experiment 1) and 104 and 136 kg
(experiment 2). Mean tissue accretion rates were
determined from the difference between final and
initial compositions, divided by the days on test.
These means then were used to test linear and
quadratic effects of digestible lysine.

72-5 to 136 kg. Linear and quadratic polynomials

(Peterson, 1985) were used to evaluate the effect of
digestible lysine. Break-point analysis (Walker and
Carmer, 1967) was used to determine the inflexion
point when the quadratic function was significant. It
provided for a 0-95 lower confidence limit which
assured that the optimal dose exceeded the lower
limit. The inflexion point for a lysine requirement
was determined for carcass protein and lipid
accretions from 34 to 72-5 kg. From 72-5 to 136 kg, the
inflexion point was not estimated for ADG and
carcass protein and lipid accretion rates because no
significant quadratic response was observed.
Body component growth curves for the treatment periods.

The procedures for body component growth curves

have been described previously by Whittemore et al.
(1988) and Schinckel (1992). Two functions, one
relating live weight to time and one relating the body
component mass to live weight, were used to
establish the body component accretion rates at each
age or weight. To determine the relationship
between live-weight gain and days on test regression
analysis was used first in the following equation to
solve for regression coefficients:
live weight = ba + b-1 (day) + b2 (day2).

Statistical procedures
Weight interval performance analyses. Analysis of

variance utilizing the GLM function of Statistical

Analysis Systems Institute (1988) was used to obtain
the least-square means for ADG and carcass CP and
lipid accretions from the periods of 34 to 72-5 kg and

An allometric equation then was used to determine

the coefficients of carcass CP and lipid accretion
relative to live weight:
Y = aXh

Friesen, Nelssen, Goodband, Tokach, Schinckel and Einstein

162

where Y = carcass CP or lipid, a = scale constant, X =

body weight, and b = relative growth coefficient. The
equation was linearized as log Y = log a + b log X to
utilize least-squares regression analyses. Component
(carcass CP or lipid) gain then was related to liveweight gain by multiplying the derivative of live
weight over time by the derivative of component
weight over live weight:
dComponent/dTime = dLive weight/dTime
X dComponent/dLive weight.
The derivative of live weight over time (daily gain at
each day) is equal to bx + 2b2 (day). The derivatives of
the carcass CP and lipid components on live weight
are equal to abXbA. The interactions between
digestible lysine and the change in BW over time
were tested. The significance of the analyses
indicated that separate regression equations were
needed for each digestible lysine treatment.

Loewer, 1986; Schinckel, 1992). This function was

fitted for two reasons over other three parameter
non-linear functions. It results in a better fit to pig
data based on residual standard deviations and
Durbin Watson statistics. Also, Monte Carlo
simulation results of a two-step solution procedure
indicated that the function was flexible yet stable,
consistently resulting in unbiased estimates of WTM,
m and a (Sun, Schinckel, Einstein, Yuan and Randin,
1993) in small data sets (no. = 16 or 24 pigs). The
following equation gives the relationship of live
weight to age of the pig:
live weight = 209-3 kg X (1 - EXP (-0-0000395
X(age126))) + l-5.
Carcass CP and lipid mass data from pigs at 34, 54,
72-5, 104, and 136 g were fitted to allometric
equations. The functions developed were:
carcass CP = (10") X (weight1 03);
carcass lipid = (10") X (weight1 536).

Overall body component growth curves. Total live

weight and carcass CP and lipid functions were

fitted from 34 to 136 kg. Gilts (three gilts per pen, six
pens per treatment) that were given 9-4 g/kg
digestible lysine in experiment 1 (34 to 72-5 kg) and
experiment 2 (72-5 to 136 kg) were used to develop
these models. Pig weights were collected weekly,
and carcass composition was determined (six gilts
per slaughter weight) at 34, 54, 72-5, 104 and 136 kg.
The function: M, = wtm (l-e""1'3) where Wti = weight
gain (kg) from birth (1-5 kg); e = mathematical
constant; m = exponential growth delay function; t =
days of age; a = kinetic order constant; wtm = weight
at maturity - birth weight; was used to fit daily
weight gain (Bridges, Turner, Smith, Stahly and

These curves were used to estimate maximum ADC

and carcass CP and lipid accretion.

Results
ADG, and CP and lipid accretion rates of gilts from
32 to 72-5 and 72-5 to 136 kg are presented in Table 3.
The carcass CP and lipid contents for gilts in
experiment 1 and 2 are presented in Table 4. These
values were used for determining their appropriate
compositional curve. The initial weights (b0) were
similar for all digestible lysine levels at 34
(experiment 1) and 72-5 kg (experiment 2; Table 5).

Table 3 The influence of digestible lysine in high-lean growth gilts on end point average daily gain (ADG) and carcass crude protein (CP)
and lipid accretion (means calculated from six observations per treatment at each body weight)
Digestible lysine (g/kg)

34 to 72-5 kg
ADG (kg)t
CP accretion (g)t^:
Lipid accretion (g)||
72-5 to 136 kg
ADG (kg)
CP accretion (g)1
Lipid accretion (g)
t
t

||
1

5-4

6-4

7-4

8-4

9.4

10-4

0-68
83
107

0-75
90
99

0-74
111
79

0-83
113
73

0-87
123
80

0-82
110
84

0-84
87
172

0-86
92
193

0-92
92
167

0-91
99
143

0-85
103
164

Linear effect of digestible lysine (P < 0-01).

Quadratic effect of digestible lysine (P < 0-05).
Linear effect of digestible lysine (P < 0-05).
Quadratic effect of digestible lysine (P < 0-10).
Linear effect of digestible lysine (P < 0-10).

s.e.
0-03
6-0
10-0
0-03
8-5
21-2

Growth modelling in pigs

163

Table 4 Carcass chemical composition (g/kg) of gilts slaughtered at 34, 54, 72.5, 104 and 136 kgi
34 kg
Digestible lysine (g/kg)
Initial
5-4
6-4
7-4
8-4
9-4
10-4

54 kg

CP

Lipid

179-0

107-0

72-5 kg

104 kg
11

CP

Lipid}:

cpt

Lipid*

171-5
176-7
184-0
179-5
178-9
178-7

146-2
126-3
113-0
117-0
113-0
98-4

1671
169-9
187-6
183-7
171-0
1810

162-3
146-8
124-5
115-3
114-0
123-7

136 kg

CP

Lipid

CP

Lipid

167-9
165-7
166-3
163-1
166-3

174-3
160-2
185-2
174-9
185-0

148-0
144-2
151-2
155-8
158-5

201-0
207-5
194-9
173-7
187-6

+ Means calculated from six pigs per treatment at each body weight. The head, leaf fat, and viscera were removed at
slaughter and not included in carcass chemical composition or determination of tissue accretion rate.
X Linear effect of digestible lysine (P < 001).
Quadratic effect of digestible lysine (P < 0-05).
|| Quadratic effect of digestible lysine (P < 0-01).

The linear (frj) and quadratic (b2) terms were different

(P< 0-001) for each digestible lysine level in both
experiments 1 and 2. The b^ terms increased as
digestible lysine level increased from 5-4 to 10-4
g/kg, resulting in greater ADG for gilts given
increased digestible lysine (Figure 1). The b2 values
were positive for gilts given 5-4, 7-4 and 8-4 g/kg,
indicating increased ADG as BW or time increased.
The negative b2 terms for gilts given 6-4, 9-4 and
10-4 g/kg suggest reduced ADG as BW increased.
For experiment 2, the bA and b2 terms were different
(P < 0-001) for each digestible lysine level. The b-j
term increased as digestible lysine increased from 5-4
to 9-4 g/kg, resulting in greater ADG (P < 0-001).
ADG (Figure 2) was reduced as BW increased, as
reflected by the negative b2 terms for each digestible
lysine level. These b2 terms were influenced
(P < 0-001) by digestible lysine for gilts given 6-4, 8-4
and 9-4 g/kg digestible lysine.
The intercepts (a) developed for the carcass CP
allometric functions from 34 to 72-5 kg decreased

(P < 0-001) as digestible lysine increased (Table 6).

The allometric growth coefficients (b) were greater
(P < 0-001) as digestible lysine increased from 5-4 to
10-4 g/kg and peaked at 7-4 g/kg. The intercepts for
carcass lipid accretion were influenced (P < 0-001) by
digestible lysine (Table 6). Carcass lipid growth
coefficients (b) were reduced (P < 0-001) as a result of
increased digestible lysine. Similarly, carcass CP and
lipid intercepts were influenced (P < 0-001) as
digestible lysine increased in experiment 2 (72-5 to
136 kg; Table 7). The allometric growth coefficients
for carcass CP increased (P < 0-001), whereas carcass
lipid coefficients decreased (P < 0-001) as digestible
lysine increased from 5-4 to 9-4 g/kg at 72-5 to
136 kg.
The data in Table 3 and Figure 1 represent weight
interval performance analyses and the corresponding
regression analyses for ADG and carcass CP and
lipid accretions from 34 to 72-5 kg. Weight interval
performance analyses for ADG (Table 3) indicated
greater (linear, P < 0-01) gains as a result of increased

Table 5 Live weight growth parameters for high-lean growth gilts fed from 34 to 72-5 kg and 72-5 to 136 kg where live weight gain on test
= b 0 + b/dni/s) + b1(days)1i
34 to 72-5 kg
Digestible lysine (g/kg)
5-4
6-4
7-4
8-4
9-4
10-4

b2

bo
34-306
33-775
34-306
34-332
33-828
34-306

72-5 to 134 kg

0-592 (0-037)***
0-760 (0-039)***
0-632 (0-039)***
0-710 (0-046)***
0-852 (0-046)***
0-995 (0-043)***

0-0009 (0-0009)
-0-00118 (0-0009)
0-00136 (0-0009)
0-00227 (0-0014)
-0-00090 (0-0014)
-0-00635 (0-0009)***

bo

bx

b2

73-702
72-568
71-838
71-411
72-142

0-900 (0-032)***
0-957 (0-032)***
0-935 (0-032)***
0-998 (0-031)***
0-963 (0-031)**"

-0-0009 (0-0005)
-0-0018 (0-0005)**
-0-0009 (00005)
-0-0023 (0-0005)***
-0-0018 (0-0005)**

+ b0 is the initial mean weight for the treatment at the beginning of the experiment.
Probability levels are for deviations of coefficients from 0.

Friesen, Nelssen, Goodband, Tokach, Schinckel and Einstein

164

1-05

1-00

0-40

34

40

46
52
58
64
Body weight (kg)

72-5 82-5 92-5 102-5 112-5 122-5

Body weight (kg)

72

136

150
(b)
^130
<s
^SpllO

-~

IX
(13

70

__ e "

_ - 9
_ ^ -^-v.
-e
^*'1\

U
1

34

40

46
52
58
Body weight (kg)

64

72

72-5 82-5 92-5 102-5 112-5 122-5

Body weight (kg)
250

(c)

^220
o

T3

S 190

g 160 h
U

u
40

46
52
58
Body weight (kg)

130
64

72

72-5 82-5 92-5 102-5 112-5 122-5

Body weight (kg)

136

Figure 1 The influence of digestible lysine (O 5-4; + 6-4;

7-4; 8-4; * 9-4; and 10-4 g/kg) in high-lean growth gilts
fed from 34 to 72-5 kg on the change in (a) average daily
gain (ADG) over time; (b) change in carcass crude protein
(CP) accretion over time; (c) the change in lipid accretion
over time.

Figure 2 The influence of digestible lysine (O 5-4; + 6-4;

7-4; a 8-4; * 9-4; g/kg) in high-lean growth gilts fed from
72-5 to 136 kg on the change in (a) average daily gain (ADG)
over time; (b) the change in carcass crude protein (CP)
accretion over time; (c) the change in lipid accretion over
time.

digestible lysine. ADG appears to be maximized for

high-lean growth gilts given 8-4 to 9-4 g total lysine
per kg (18 g digestible lysine per day or 22 g
digestible lysine per day intake). Live-weight growth
curve analyses of ADG (Figure 3a) indicated
maximum ADG (P < 0-001) for high-lean growth gilts
given 10-4, 9-4 and 8-4 g/kg digestible lysine from 34
to 44 kg, 44 to 54 kg, and 54 to 72-5 kg, respectively.
These diets would provide 16,18, and 18 g digestible
lysine per day intake (19, 21, and 21 g total lysine per

day intake, respectively) for the specified weight

periods. Weight interval performance analyses
indicated increased (quadratic, P < 0-05) CP accretion
as digestible lysine increased (Table 3). Inflexion
point analysis projected the maximum response at
7-9 g digestible lysine per kg (14 g/day) (9-4 g/kg or
17 g total lysine per day). Regression analyses
(Figure 3b) showed that carcass CP accretion was
maximum (P < 0-001) for gilts given 10-4, 9-4, and
8-4 g/kg digestible lysine from 34 to 44 kg, 44 to

Growth modelling in pigs

165

Table 6 Weight of carcass crude protein and lipid components using the relationship o/Y = aXb, where Y is the component and X the live
weight (kg), for high-lean growth gilts fed from 34 to 72-5 kgf
Carcass crude protein
Digestible lysine (g/kg)
5-4
6-4
7-4
8-4
9-4
10-4

Carcass lipid
2

0-0938
0-0820
0-0540
0-0600
0-0610
0-0643

1-052(0-039)
1-089(0-041)
1-209(0-059)**
1-178(0-042)***
1-177(0-046)***
1-157(0-056)**

0-98
0-98
0-96
0-98
0-97
0-96

a
0-0049
0-0104
0-0182
0-0286
0-0268
0-0275

b
1-742(0-178)***
1-526(0-156)**
1-364(0-129)**
1-240 (0-154)
1-263 (0-138)
1-240(0-140)

R2
0-87
0-87
0-87
0-79
0-83
0-82

+ Probability levels are for deviations of coefficients from 1-00.

54 kg, and 54 to 72-5 kg respectively. The analysis of

CP accretion would result in lysine estimates
(digestible and total) to those from the ADG analysis.
Carcass lipid accretion decreased (quadratic,
P < 0-10) as digestible lysine increased (Table 3). The
inflexion point for minimum carcass lipid gain was
calculated at 7-1 g digestible lysine per kg (13 g/day)
(8-4 g/kg or 15 g total lysine per day). The regression
of carcass lipid accretion over BW indicated a linear
increase (P < 0-001) in lipid gain for all dietary
treatments except in gilts given 10-4 g/kg (Figure 3c).

CP accretion over BW (Figure 3b) was maximum

(P < 0-001) for gilts given 9-4 g digestible lysine per
kg (28 g/day) or 11-2 g total lysine per kg (33 g/day).
However, as BW increased, carcass CP accretion
continually decreased (P < 0-001). Carcass lipid
accretion (Table 3) was not influenced (P < 0-10) by
digestible lysine in high-lean growth gilts fed from
72-5 to 136 kg. The regression analyses of carcass
lipid accretion (Figure 3c) indicated increased lipid
accretion for gilts given all diets except that with
8-4 g digestible lysine per kg.

Weight interval peformance analyses from 72-5 to

136 kg indicated that digestible lysine did not
influence significantly (P > 010) ADG (Table 3).
Numerically, ADG appeared to be maximized for
gilts given 7-4 g/kg (22 g digestible lysine per day) or
8-8 g total lysine per kg (27 g/day). Regression
analyses of BW gain over time (Figure 2a) indicated
maximum (P < 0-001) ADG for gilts given 8-4 and
7-4 g digestible lysine per kg from 72-5 to 92-5 kg and
92-5 to 136 kg, respectively. These estimates would
provide 25 and 22 g digestible lysine per day intake).
Carcass CP accretion (Table 3) was greater (linear,
P < 0-10) as digestible lysine increased. This increase
in CP accretion would require 9-4 g digestible lysine
per kg (28 g/day) or 11-2 g total lysine per kg
(33 g/day). Regression analyses showed that carcass

ADG (Figure 3a) and carcass CP accretion (Figure 3b)

increased at a decreasing rate from 38 kg to about 98
and 100 kg, respectively. ADG (0-85 kg/day) and
carcass CP accretion (107 g/day) were maximized at
948 kg body weight. Carcass lipid accretion (Figure
3c) increased linearly (P < 0-001) as BW increased.
The rate of carcass CP and lipid accretion decreased
as BW increased.

Discussion
Whittemore et al. (1988) indicated that fitting liveweight gain on test to a Gompertz function and
allometric equations was the most accurate way to
determine the rate of protein deposition. Unbiased
comparisons of carcass tissue growth using

Table 7 Weight of carcass crude protein and lipid components using the relationship ofY = aXb, where Y is the component and X the live
weight (kg), for high-lean growth gilts fed from 72-5 to 136 kgf
Carcass crude protein
Digestible lysine (g/kg)
5-4
6-4
7-4
8-4
9-4

Carcass lipid

R2

0-4410
0-3218
0-3589
0-2828
0-2378

0-718 (0-058)***
0-788 (0-064)**
0-763 (0-047)***
0-817(0-091)
0-857 (0-050)**

0-91
0-90
0-94
0-85
0-95

0-0021**
0-0010**
0-0024**
0-0061**"
0-0024**"

t Probability levels of deviations of coefficients from 1 -00.

b
1-876 (0-178)***
2-030 (0-152)***
1-843(0-198)***
1-622(0-179)**
1-836(0-179)***

R2
0-87
0-91
0-85
0-84
0-87

Friesen, Nelssen, Goodband, Tokach, Schinckel and Einstein

166

portions of the curve were completed by each

experiment. Two experiments were conducted to
avoid confounding the data set from 72-5 to 136 kg
by feeding deficient diets from 34 to 72-5 kg.

0-88

Fitting a non-linear (age and live-weight

relationship) function (Bridges et al., 1986) results in
shapes of the total BW and carcass CP accretion
curves that are slightly different compared with
those derived from the use of the polynomial (linear
and quadratic) functions fitted for each individual
experiment. The predicted CP accretion over time is
highly sensitive to the function being fitted
(Schinckel, 1994). The sigmoidal (non-linear) growth
function fits a line assuming that live-weight gain
increases to an inflexion point, reaches a plateau, and
then decreases. However, the function is fitted
under the assumption that maximum growth
performance is achieved throughout the entire
growth period (Whittemore, 1986; Whittemore et al.,
1988). However, the performance recently described
in segregated early weaned pigs indicates that
performance is not maximized in traditional 21- or
28-day weaning (Dritz, Nelssen, Goodband, and
Tokach, 1994). This type of function could not be
fitted to the individual experiments, because weight
intervals of 38 and 61 kg for experiments 1 and 2,
respectively, were analysed instead of the entire
growth period. Thus, the polynomial functions will
give a different representation of the data, which is
more accurate because it does not assume the
sigmoidal shape as does the age-dependent function
(Bridges et al, 1986).

78
Body weight (kg)

58

78
98
Body weight (kg)

118

136

Figure 3 The predicted change in (a) average daily gain

(ADG); (b), carcass crude protein (CP) accretion rate; (c)
carcas lipid accretion rate for high-lean growth gilts fed
94 g/kg digestible lysine from 34 to 136 kg.

allometric equations were conducted by Gu,

Schinckel and Martin 1992) for five different
genotypes. The data in our experiment (34 to 72-5 kg
and 72-5 to 136 kg) could not be fitted to a sigmoidshaped function, because only short weight periods
were analysed instead of the entire growth curve.
The entire growth curve is needed to use the
sigmoid-shape growth equation so that an inflexion
point is clearly defined (Schinckel, 1992 and 1994).
Live-weight gain on test in our experiment was
determined by polynomial regression, because two

The ADG and carcass CP gain data from 34 to

72-5 kg are in agreement with the linear-plateau
portion of the growth curve proposed by
Whittemore (1986). Weight interval performance
analyses fit a linear response to increased digestible
lysine. These data are similar to the linear-plateau
relationship for protein deposition in the growing
boar (Campbell and King, 1982; Campbell, Taverner
and Curie, 1984) and castrated male (Oyeleke,
Balogun, Fetuga and Babatunde, 1988). Campbell,
Taverner and Curie (1988) suggested that the dietary
protein content could be decreased in both gilts and
boars as end-point weight increased from 50 to 90 kg.
However, these end-point analyses and simple
regression models do not characterize the changing
response to dietary lysine as BW increases. The nonlinear regression models developed in our
experiment indicate that the response to digestible
lysine changed as BW increased from 34 to 72-5 kg.
At 34 kg, food intake (i.e. lysine intake) was limited
by reduced appetite in the growing pig (Rao and
McCracken, 1992) Thus, each incremental increase in
digestible lysine resulted in greater total weight and

Growth modelling in pigs

carcass CP gain. Lipid accretion, on the other hand,
decreased as a result of greater lysine intake. This
shift in the composition of gain to greater CP
accretion suggests that digestible lysine was the
limiting factor for CP deposition (Campbell et al.,
1988). Campbell et al. (1984) observed a linear
decrease in empty body protein gain as protein
intake was increased from 95 to 186 g/kg for boars
fed from 45 to 90 kg. However, these authors
proposed that protein retention is dependent on level
of feeding (energy intake) and not on protein intake
during the linear portion of the protein deposition
curve. These concepts are further supported by the
results of Rao and McCracken (1992) in growing
boars (33 to 90 kg), which indicated greater ADG,
empty body protein gain, and efficiency of gain as
energy intake increased from 14-8 to 15-2 MJ ME per
kg. In our experiment, the dietary ME was held
constant at 14-3 MJ/kg. Thus, the resulting increase
in carcass protein gain was dependent on increased
lysine intake.
The regression data for carcass CP accretion suggest
that gilts given 6-4 g digestible lysine per kg
exhibited proportionately only 0-65 of the maximal
protein with gilts given 10-4 g digestible lysine per
kg. The difference in the regression models
represents the influence of digestible lysine on BW
where peak protein gain was attained. De Greef and
Verstegen (1993) suggested that previous nutrient
intake will influence the shape of compositional
curves, offering the potential for compensatory
protein gain (Critser, Miller, Lewis and Wolverton,
1993; Whang and Easter, 1994). Our data suggest that
the 7-4 g digestible lysine per kg diet given prior to
the first experiment limited protein deposition. Thus,
feeding greater levels of digestible lysine resulted in
larger increases in protein gain, potentially as
compensatory gain as well as meeting the lysine
need for maximum carcass CP gain.
Gilts given 10-4 g digestible lysine per kg had
maximum protein gain at the onset of the
experiment. However, the rapid reduction in CP gain
for those gilts is rather alarming. Potentially, CP gain
could be decreased by increased amino acid
catabolism, resulting in a reduction in the net energy
for gain (Holmes, Carr and Pearson, 1980; Campbell
and King, 1982). However, the formation of urea
from the catabolism of excess amino acids results in
only a two-ATP depletion per molecule of amino
acid (Stryer, 1988). A second rationale is
compensatory protein gain for the gilts given 7-4, 8-4,
and 9-4 g digestible lysine per kg (Whang and Easter,
1994). Maximum protein gain was attained at
different BWs for gilts receiving various levels of
digestible lysine. Thus, reduced digestible lysine
(104 i'. 8-4 g/kg) resulted in heavier BWs for

167

attaining maximum protein gain. These data

illustrate two important concepts. First, the influence
of under-feeding lysine during the growing phase,
i.e. in reduced carcass protein gain. This concept was
characterized by Stahly, Cromwell and Terhune
(1988), who reported the dramatic reduction in lean
growth for high-lean growth barrows when given
lysine-limiting diets from 50 to 110 kg. Secondly, the
regression analysis emphasizes the importance of
phase feeding throughout the linear portion of the
growth curve (34 to 72-5 kg).
Lipid accretion from 34 to 72-5 kg increased as BW
increased. Similarly, Shields, Mahan and Graham
(1983) suggested a linear increase in lipid gain as a
result of increased BW. Lipid accretion increased to a
lesser extent for gilts given 8-4 to 9-4 g digestible
lysine per kg than for gilts given a deficient diet (5-4
to 6-4 g digestible lysine per kg). The resulting
increase in amino acid intake is equated with greater
CP gain rather than lipid gain (de Greef and
Verstegen, 1993). Thus, carcass lipid gain is
minimized when CP gain is maximum in high-lean
growth gilts. In the grower phase, lysine will be used
preferentially for protein synthesis (Finkelstein,
1990). However, not all protein synthesis is
associated with lean tissue gain (Etherton,
Wangsness, Hammers and Ziegler, 1974). The high
impetus muscles of the ham and loin will respond to
increased lysine intake to a greater extent in the
grower period (34 to 72-5 kg) than in the finishing
period (72-5 to 136 kg). But the increase in lipid gain
is consistent with protein gain for the adipose tissue
matrix, and minimal lipid is needed for body
maintenance (Whittemore, 1986).
The gilts in experiment 2 (72-5 to 136 kg) did not
exhibit the magnitude of response to digestible lysine
seen in experiment 1. Generally, the ADG and
carcass CP gain decreased linearly, whereas carcass
lipid gain increased linearly as a result of greater
BW. These results are consistent with previous
reports of Shields et al. (1983). The decreasing rate of
total body and carcass CP gain indicated that the
gilts in this experiment were beyond the inflexion
point for maximum weight gain (Whittemore and
Fawcett, 1976). Similarly, Shoup (1991) and Gu et al.
(1992) indicated a linear reduction in lean tissue gain
as BW increased from 60 to 130 kg. The resulting
increase in food intake with increased BW was
attributed to greater lipid deposition. These results
are similar to the increased lipid gain with increased
energy intake in growing boars from 30 to 90 kg by
Campbell and Taverner (1988). The efficiency of
lysine utilization appears to be reduced for heavier
gilts. The total lysine intake in finisher gilts was
5g/day greater to support carcass CP deposition
rates that were 20 to 30 g/day less than those of

168

Friesen, Nelssen, Goodband, Tokach, Schinckel and Einstein

grower gilts. This was apparent in reduced food

efficiencies as BW increased (Shields et ah, 1983). The
poorer efficiencies for lysine use can be attributed
further to increase lipid deposition that resulted from
the greater food intake and excess nutrient intake
(Campbell et ah, 1984).
Thus, performance and carcass lean deposition
apparently are reduced as a result of increased BW.
Our data suggest that greater digestible lysine intake
increased CP gain and decreased lipid gain.
However, the efficiency of lysine utilization in
experiment 2 was reduced compared with that of
gilts in experiment 1. This reduction was evident by
the decreased slope of carcass CP on lysine intake.
These data are in agreement with de Greef and
Verstegen (1993), who reported increased carcass
leanness with increasing energy intake combined
with sufficient amino acid intakes. However, the
increased protein gain was proportionally greater
than the non-lean tissue gain (i.e. lipid gain), which
is evident in our data. Although carcass CP gain is
greater with increased digestible lysine the cost of
achieving maximum protein gain is not economically
feasible (Crenshaw, Gahl and Benevenga, 1994).
Therefore, economics will determine the level of
lysine that can be offered for maximum profit.
In conclusion, the growth and composition models
described in this experiment indicate the importance
of regression modelling to describe accurately
nutrient requirements in growing-finishing pigs. In
the growing period (34 to 72-5 kg), the models
accurately estimated digestible lysine requirement
for maximum carcass protein gain as BW increased.
Although digestible lysine caused less of a response
in the finishing period (72-5 to 136 kg), the models
characterized the diminishing response to digestible
lysine as BW increased. Thus, these models provide
basic information that can be incorporated with
economic considerations to maximize profit.

Acknowledgements
Contribution no. 95-63-J from the Kansas Agriculture
Experiment Station, Manhattan, 66506.

References
Association of Official Analytical Chemists. 1990. Official
methods of analysis. 15th ed. Association of Official
Analytical Chemists, Arlington, Va.
Agricultural Research Council. 1981. The nutrient
requirements of pigs. Commonwealth Agricultural Bureaux,
Slough.
Baker, D. H. 1986. Critical review: problems and pitfalls in
animal experiments designed to establish dietary
requirements for essential nutrients, journal of Nutrition 116:
2339.

Benevenga, N. J., Gahl, M. J., Crenshaw, T. D. and Finke,

M. D. 1994. Protein and amino acid requirements for
maintenance and amino acid requirements for growth of
laboratory rats. Journal of Nutrition 124: 451.
Bridges, T. C, Turner, U. W., Smith, E. M., Stahly, T. S.
and Loewer, O. J. 1986. A mathematical procedure for
estimating animal growth and body composition.
Transactions of the Society for Agricultural Engineering 29:
1342-1347.
Campbell, R. G. and King, R. H. 1982. The influence of
dietary protein and level of feeding on the growth
performance and carcass characteristics of entire and
castrated male pigs. Animal Production 35: 172-184.
Campbell, R. G. and Taverner, M. R. 1988. Genotype and
sex effects on the relationship between energy intake and
protein deposition in growing pigs, journal of Animal Science
66: 676-686.
Campbell, R. G., Taverner, M. R. and Curie, D. M. 1984.
Effect of feeding level and dietary protein content on the
growth, body composition and rate of protein deposition in
pigs growing from 45 to 90 kg. Animal Production 38:
233-240.
Campbell, R. G., Taverner, M. R. and Curie, D. M. 1988.
The effects of sex and live weight on the growing pigs'
response to dietary protein. Animal Production 46:123-130.
Chung, T. K. and Baker, D. H. 1992. Ideal amino acid
pattern for 10-kilogram pigs, journal of Animal Science 70:
3102.
Cook, D. A. 1991. The conceptual analysis of a dynamic
mathematical model for the estimation of the amino acid
requirements for pigs from weaning to maturity. Ph. D.
dissertation, University of Illinois, Urbana-Champaign.
Crenshaw, T. D., Gahl, M. J. and Benevenga, N. J. 1994.
The impact of diminishing returns on performance and feed
costs for growing and finishing pigs, journal of Animal
Science 72: sttppl. I, p. 59 (abstr.).
Critser, D. J., Miller, P. S., Lewis, A. J. and Wolverton,
C. K. 1993. The effects of dietary protein concentration
during realimentation on compensatory growth in barrows
and gilts, journal of Animal Science 71: suppl. 1, p. 178
(abstr.).
Dritz, S. S., Nelssen, J. L., Goodband, R. D. and Tokach,
M. D. 1994. Application of segregated early weaning
technology in the commercial swine industry. Compendium
on Continuing Education for the Practicing Veterinarian 16: 677.
Etherton, T. D., Wangsness, P. J., Hammers, V. M. and
Ziegler, J. H. 1974. Effect of dietary restriction on carcass
composition and adipocyte cellularity of swine with
different propensities for obesity. Journal of Nutrition 112:
2314.
Finkelstein, J. D. 1990. Methionine metabolism in
mammals, journal of Nutritional Biochemistry 1: 228.
Friesen, K. G., Nelssen, J. L., Goodband, R. D., Tokach,
M. D., Unruh, J. A., Kropf, D. H. and Kerr, B. J. 1995a.
Influence of dietary lysine on growth and carcass
composition of high-lean growth gilts fed from 34 to 72
kilograms, journal of Animal Science. 72: 1761.
Friesen, K. G., Nelssen, J. L., Goodband, R. D., Tokach,
M. D., Unruh, J. A., Kropf, D. H. and Kerr, B. J. 1995b. The
effect of dietary lysine on growth and carcass composition

Growth modelling in pigs

169

in high-lean growth gilts fed from 72 to 136 kilograms.

postweaning growth and carcass merit and the
characterization of three genotypes of swine. M. S. thesis,
journal of Animal Science. In press.
Purdue University, West Lafayette, IN.
Greef, K. H. de and Verstegen, M. W. A. 1993. Partitioning
Stahly, T. S., Cromwell, G. L. and Terhune, D. 1988.
of protein and lipid deposition in the body of growing pigs.
Response of pigs from high and low growth genotypes to
Livestock Production Science 35: 317.
dietary lysine levels. Journal of Animal Science 66: suppl. 1,
Gu, Y., Schinckel, A. P. and Martin, T. G. 1992. Growth,
p. 137 (abstr.).
development, and carcass composition in five genotypes of
Statistical Analysis Systems Institute. 1988. SAS/STAT
swine. Journal of Animal Science 70:1719.
user's guide. Release 6.03. Statistical Analysis Systems
Holmes, C. W., Carr, J. R. and Pearson, G. 1980. Some
Institute Inc., Cary, NC.
aspects of the energy and nitrogen metabolism of boars,
Stryer, L. 1988. Biochemistry. 3rd ed. Freeman, New York.
gilts and barrows given diets containing different
concentrations of protein. Animal Production 31: 279-289.
Sun, F., Schinckel, A., Einstein, M., Yuan, L. and Randin,
Knabe, D. A., LaRue, D. C, Gregg, E. J., Martinez, G. M.
R. 1993. Solution and testing of three parameter live weight
and Tanksley, T. D. 1989. Apparent digestiblity of nitrogen
growth curve in a two stage procedure. Proceedings of the
and amino acids in protein feedstuffs by growing pigs.
cooperative lean growth update, Purdue University.
Journal of Animal Science 67: 441.
Walker, W. M. and Carmer, S. G. 1967. Determination of
National Research Council. 1988. Nutrient requirements of input levels for a selected probability of response in a
curvilinear regression function. Agronomy Journal 59:161.
swine. 9th ed. National Academy Press, Washington, DC.
Wang, T. C. and Fuller, M. F. 1989. The optimum dietary
Oyeleke, M. O., Balogun, O. O., Fetuga, B. L. and
amino acid pattern for growing pigs. British Journal of
Babatunde, G. M. 1988. Influence of dietary protein levels
Nutrition 62: 77.
on rate of tissue deposition and individual muscle
development of growing European pigs in a tropical
Watt, D. L., DeShazer, J. A., Ewan, R. C, Harrold, R. L.,
environment. Journal of Agricultural Science, Cambridge 110:Mahan, D. D. and Schwab, G. D. 1987. NCCISWINE:
377.
Housing, nutrition, and growth simulation model. Applied
Peterson, R. G. 1985. Design and analysis of experiments. Agricultural Research 2: 218.
Marcel Dekker, New York.
Whang, K. E. and Easter, R. A. 1994. Effect of starter
feeding program on growth performance and protein gain
Rao, D. S. and McCracken, K. J. 1992. Energy:protein
from weaning to market weight in barrows and gilts.
interactions in growing boars of high genetic potential for
Journal of Animal Science 72: suppl. 1, p. 65 (abstr.).
lean growth. 1. Effects on growth, carcass characteristics
and organ weights. Animal Production 54: 75-82.
Whittemore, C. T. 1986. An approach to pig growth
modeling. Journal of Animal Science 63: 615.
Schinckel, A. P. 1992. Concepts of lean growth modeling
and methods of describing genetic differences for
Whittemore, C. T. and Fawcett, R. H. 1976. Theoretical
application to lean growth models. Proceedings of the aspects of a flexible model to simulate protein and lipid
National Pork Producers Council lean-growth modeling growth in pigs. Animal Production 22: 87-96.
symposium, Dcs Moines, IA, pp. 53-73.
Whittemore, C. T., Tullis, J. B. and Emmans, G. C. 1988.
Schinckel, A. P. 1994. Nutrient requirements of modern pig
Protein growth in pigs. Animal Production 46: 437-445.
genotypes. In Recent advances in animal nutrition (ed. P. C.
Williams,
N. H., Stahly, T. S. and Zimmerman, D. R. 1994.
Garnsworthy and D. J. A. Cole). Nottingham Press,
Impact of immune system activation on growth and amino
Loughborough.
acid needs of pigs from 6 to 114 kg body weight. Journal of
Shields, R. G., Mahan, D. C. and Graham, P. L. 1983.
Animal Science 72: suppl. I, p. 57 (abstr.).
Changes in swine body composition from birth to 145 kg.
Journal of Animal Science 57: 43.
Shoup, M. A. 1991. The effects of recombinant porcine
(Received 24 July 1995Accepted 25 August 1995)
somatotropin (pST), slaughter weight and genotype on