By GLEN BRODERICK, ESSI EVANS
and BRITTANY DYCK*
EDITOR’S NOTE: The authors re-
cently took part in a fact-ﬁ nding
mission and visited several feed
mills in Wisconsin. One important fact
that came to light was the difﬁ culty in
determining how to place a value on pro-
tein ingredients. Many purchasers used
a dollar spread to assess the value of a
protein, and others used a ratio system.
The discussions that ensued revealed a
number of pitfalls in these methods.
ONE of the toughest economic deci-
sions for nutritionists to make is select-
ing dietary ingredients, as each ingredi-
ent contributes more than one nutrient.
For example, barley may have less en-
ergy than corn grain, but it also provides
Then, there are ruminal effects: Barley,
which contains less total starch than
corn, may contribute more rumen-avail-
able starch. This becomes even more
complex when proteins are compared.
Should the ingredients be valued on
the basis of cost per ton, cost per unit
of crude protein (CP) or cost per unit of
metabolizable protein (MP)? What about
amino acid proﬁ les of ingredients?
Selecting the wrong metric can make
a considerable difference to the bot-
tom line. The purpose of this article is
to demonstrate that all of these may
be misleading if the value to the ani-
mal is not taken into account, as well
as to provide a more accurate means
of estimating the value of protein in-
gredients in such a way that the pur-
chaser is less vulnerable to mistakes.
Cost per ton of CP
Having data only on the CP content of an
unknown feedstuff tells little about its
MP and amino acid contents. The func-
Feedstuffs, June 5, 2017
© 2017 Feedstuffs. Reprinted with permission from Vol. 89, No.06, June 5, 2017
Comparison of feed proteins for
dairy cows takes careful thought
Nutritionists struggle to estimate the value of protein
ingredients, and selecting the wrong metric can make a
considerable difference to the bottom line.
*Glen Broderick is the president of
Broderick Nutrition & Research LLC,
Madison, Wis. Essi Evans is the president
of Technical Advisory Services Inc.,
Bowmanville, Ont. Brittany Dyck is the
meal manager for the Canola Council of
Canada, Winnipeg, Man.
tion of dietary CP is to supply the cow
with MP as absorbed amino acids, but
any oversupply of dietary CP that does
not contribute to absorbed amino ac-
ids that are used for production will be
largely lost in the urine. Urinary nitrogen
is the most polluting form of excretory
1. Effects of increasing dietary crude protein on milk
production and composition and nitrogen metabolism
parameters (Olmos Colmenero and Broderick, 2006b)
---------------------Dietary CP, % of dry matter---------------------
Milk parameters 13.5 15.0 16.5 17.9 19.4
Dry matter intake, lb./day 47.6b 48.1a,b 49.6a 47.6b 47.8a,b
Bodyweight gain, lb./day 0.49 0.49 0.70 0.57 0.64
Milk yield, lb./day 80.0b 82.0a,b 84.4a 80.7b 81.6a,b
Milk fat, % 3.17c 3.26a,b,c 3.23b,c 3.49a 3.45a,b
Milk fat yield, lb./day 2.51 2.65 2.73 2.71 2.73
3.5% fat-corrected milk, lb./day 75.2b 78.5a,b 80.9a 78.7a,b 79.6a,b
Milk protein, % 3.09 3.15 3.09 3.18 3.16
Milk protein yield, lb./day 2.43b 2.54a,b 2.60a 2.49a,b 2.54a,b
Milk N:N intake 0.37a 0.34b 0.31c 0.28d 0.26e
Urea nitrogen excretion, g/day 63.2e 91.0d 128.4c 174.0b 208.1a
a,b,c,d,eMeans in rows without common superscripts are different (P < 0.05).
2. Comparison of protein sources on milk production
(Brito and Broderick, 2007)
--------------------------Added protein source--------------------------
Diet parameters Urea SBM Cottonseed meal CM
Dry matter, % 1.9 12.1 14.1 16.5
Microbial protein, g/day 2,340 2,710 2,710 2,780
RUP, g/day 540 990 1,350 1,150
Protein entering intestine, g/day 2,880 3,700 4,060 3,930
Dry matter intake, lb./day 48.7 54.3 54.5 54.9
Milk yield, lb./day 72.5 88.2 89.3 90.6
Protein yield, lb./day 2.03 2.71 2.60 2.80
Fat yield, lb./day 2.23 2.69 2.60 2.84
3. Calculated effective protein degradation, RUP and
RUP contributed by meals*
Nitrogen Effective protein RUP, % Protein, % RUP, %
solubility, % degradation, % of protein of meal of meal
CM (rapeseed meal) 20 44 56 36.9 20.6
Flax (linseed meal) 59 46 54 26.8 14.5
Lupins 80 56 44 33.8 14.9
Peas 78 71 29 25.0 7.25
SBM 17 73 27 50.6 13.7
Wheat distillers grains 24 79 21 37.5 7.9
*Hedqvist and Uden, 2006.
4. Calculation of available RUP for three ingredients*
CP, RUP (inhibitor Intestinal Available
% of method), digestibility, RUP,
Ingredient DM % of CP % of RUP % of DM
Solvent-extracted SBM 52.8 34.9 89.8 16.5
Expeller SBM 46.9 59.0 88.0 24.2
Solvent-extracted CM 41.6 44.4 76.9 14.2
*Broderick et al., 2016; AminoDat 5.0.
Feedstuffs, June 5, 2017
Numerous studies show that dairy
cows utilize feed CP (nitrogen x 6.25)
much more efﬁ ciently than other rumi-
nant livestock; however, dairy cows still
excrete about two to three times more
nitrogen in manure than in milk, as the
efﬁ ciency values in Table 1 clearly indi-
cate. Supplemental protein sources need
to complement microbial protein in or-
der for cows to be most efﬁ cient.
Cost per ton of MP, amino acids
Microbial protein is “high” quality be-
cause the essential amino acids are pres-
ent in proportions very close to those
required by the cow for the synthesis of
milk proteins. With cows milking more
than ever before, the source and quality
of RUP is becoming as important as the
amount being fed. In the past, protein
quality (the essential amino acid compo-
sition of feed protein) has been consid-
ered much less important to ruminants
because rumen microbes synthesize
high-quality protein out of the lower-
quality protein and the non-protein ni-
trogen in the diet.
So, what has happened to change this
view? Research has shown that, even for
ruminants such as dairy cows, not all
RUP is created equally. The differences
have to do with essential amino acids.
Many of the studies conducted at the
Dairy Forage Research Center compared
solvent SBM with expeller SBM and
roasted soybeans, both of which have
more RUP due to heat treatment. The
expeller SBM had 91% more RUP than
the solvent SBM. In three separate trials,
when these value-added soybean prod-
ucts were fed to supply equal amounts
of CP in the ration, the boost in milk pro-
tein for expeller SBM averaged only 54%
— a substantial improvement, but far
from the 91% increase in RUP.
In addition, similar trials were per-
formed with ﬁ sh meal. The low-soluble,
ruminant-grade ﬁ sh meal had twice as
much RUP as solvent SBM. When cows
were fed this source, their milk protein
yield response was also double that of
cows fed the solvent SBM (Broderick,
1992). Protein utilization was better with
ﬁ sh meal because of differences in the
protein quality of ﬁ sh meal compared to
This difference is because ﬁ sh meal is
much higher in methionine — one of the
essential amino acids and the one most
often ﬁ rst limiting for milk production.
It was concluded that utilization of soy-
bean RUP is limited by inadequate me-
Ruminant-grade ﬁ sh meal is quite ex-
pensive. Often, the cost is not covered
by the value of its greater protein qual-
ity. However, there are higher-quality
plant proteins that can be used. A recent
study at the Dairy Forage Research Cen-
ter compared four diets based on typi-
cal Midwest feeds (Brito and Broderick,
2007). The diets all contained about
16.5% CP, but the supplemental protein
came from four different sources (Table
5. Calculation of available RUP for three ingredients*
Available Amino Intestinal Available
RUP, Amino acid, % digestibility, amino acid,
Ingredient % of DM acid** of RUP % of RUP g/kg
Solvent-extracted SBM 16.5 Lys 6.17 90 9.16
Met 1.38 92 2.10
Met+Cys 2.84 88 4.13
His 2.63 91 3.95
Expeller SBM 24.2 Lys 6.04 89 13.00
Met 1.34 90 2.92
Met+Cys 2.83 83 5.68
His 2.61 90 5.68
Solvent-extracted CM 14.2 Lys 5.58 73 5.82
Met 1.95 84 2.33
Met+Cys 4.34 77 4.75
His 2.72 81 3.13
*Broderick et al., 2016; AminoDat 5.0.
**Lys = lysine, Met = methionine, Met + Cys = methionine + cysteine, His = histidine.
6. Results from seven studies that compared soybean
meal (SBM) and canola meal (CM)
---Milk yield- -- -Milk protein yield- --Milk fat yield--
Experiment SBM CM SBM CM SBM CM
Brito and Broderick, 2007 88.2 90.6 2.71 2.80 2.69 2.84
Broderick et al., 2015 84.5 85.8 2.81 2.86 3.52 3.65
Contreras-Govea et al., 2013 80.0 86.2 2.53 2.76 3.00 3.18
Faciola and Broderick, 2013 80.2 82.2 2.42 2.47 3.20 3.22
Broderick and Faciola, 2014 84.3 85.5 2.81 2.85 3.52 3.64
Paula et al., 2015 88.2 91.1 2.76 2.76 3.51 3.62
Moore and Kalscheur, 2016 112.9 122.8 — — — —
Averages 88.3 92.0 2.67 2.75 3.24 3.36
nitrogen, because much of it is lost as at-
mospheric ammonia or runs into surface
and ground water. In addition, the inef-
ﬁ cient use of dietary CP increases milk
Based on numerous studies conduct-
ed at the Dairy Forage Research Center
in Madison, Wis., some general conclu-
sions can be made about CP. There were
no increases in yield of milk, fat-correct-
ed milk or milk protein with more than
16.5% dietary CP. Lowering protein and
adding rumen undegradable protein
(RUP) to the diet is highly dependent on
the source of the RUP.
In one trial (Olmos Colmenero and
Broderick, 2006a), reducing CP to 15.6%
but adding RUP as heated soybean meal
(SBM) did not support production equal
to 16.5% CP. However, canola meal (CM)
was found to be a more effective source
of MP than SBM or cottonseed meal (Bri-
to et al., 2007; Broderick et al., 2015).
Supplementing rumen-protected me-
thionine has also been shown to be ef-
fective for allowing some reduction in
dietary CP without losing milk yield, par-
ticularly when methionine-poor ingredi-
ents such as SBM form the main source
of supplemental protein (Broderick et
al., 2009; Chen et al., 2011).
Studies conducted at the Dairy For-
age Research Center have demonstrated
that there are no interactions between
energy density or neutral detergent ﬁ ber
(NDF) level and response to dietary pro-
tein sources. Milk yield was increased
by reducing forage from 75% to 62% and
then to 50% of dietary dry matter, re-
sulting in diets with 36%, 32% and 28%
NDF; dietary CP was fed at about 15.1%,
16.7% and 18.4% of dry matter at each
NDF level (Broderick, 2003). The cows
responded to CP the same way at all
three energy levels. Milk and protein
yield both increased with the ﬁ rst CP in-
crement, but there was no difference be-
tween production at 16.7% and 18.4% CP.
Thus, proteins can be reviewed without
being overly concerned about energy in-
This experiment was followed up by a
study that evaluated stepwise increases
of 1.5 percentage units from 13.5% to
19.4% CP in 50% forage diets (Olmos
Colmenero and Broderick, 2006b). As
expected, milk urea nitrogen (MUN),
urinary urea and the milk nitrogen-to-
nitrogen intake (N:N) ratio reﬂ ected the
linear decline in nitrogen efﬁ ciency with
increasing CP (Table 1). Milk production
was highest on the 16.5% CP diet, and a
quadratic response was observed that
indicated that milk and protein yields
were greatest at 16.7% and 17.1% CP, re-
spectively. Overfeeding protein actually
appeared to suppress production.
Ruminants make efﬁ cient use of diets
that are poor in protein content or qual-
ity because rumen microbes synthesize
high-quality protein by capturing recy-
cled urea nitrogen that would otherwise
be excreted in the urine.
Feedstuffs, June 5, 2017
The results obtained in this study can-
not be explained by the RUP supply or
total protein ﬂ ow to the intestines. The
cows receiving cottonseed meal had the
most rumen protein outﬂ ow, but they
didn’t make as much milk or milk protein
(or fat) as the cows fed CM. The reason
likely was because CM has substantially
more methionine and better amino acid
quality than either SBM or cottonseed
Even more support was provided by
Swedish researchers Hedqvist and Uden
(2006). Their research showed that there
was a low correlation between soluble
protein and degraded protein. CM is high
in soluble protein but also high in RUP
With this information at hand, how
would one go about determining the val-
ue of available protein? Clearly, if using
CP to purchase meal, a particular protein
source might easily be overvalued or un-
Due to the lack of a relationship be-
tween protein solubility and the RUP that
appears at the duodenum, a method for
determining protein degradation from
in vitro amino acid and peptide release
was developed (Colombini et al., 2011).
This method allows users to calculate a
more accurate value for RUP and shows
that the RUP for solvent-extracted SBM,
expeller SBM and solvent-extracted CM
is 34.9%, 59.0% and 44.4% of the protein,
What is not immediately obvious is
the fact that the digestibility of the RUP
fraction can be different for different in-
gredients. Such values are not frequently
determined for ingredients used in feed-
ing dairy cattle.
However, intestinal digestibility values
have been determined on a routine ba-
sis for swine. One excellent source is a
program produced by Evonik Industries
(AminoDat 5.0). Using these values for
illustration purposes, Table 4 provides
estimations of absorbable RUP for these
Using this method, it can be seen from
Table 5 that expeller SBM provides 1.46
times per pound as much intestinally
available RUP as solvent-extracted SBM,
while solvent-extracted CM provides
0.86 as much intestinally available RUP.
These values could, therefore, be used to
estimate the worth of ingredients. If SBM
is worth $300 per ton, then expeller SBM
would be worth $438 per ton, and CM
would be valued at $258 per ton.
How about the amino acid proﬁ le? The
same method can be used to calculate
the relative value of meals using amino
As Table 5 shows, there are clear dif-
ferences in the contribution of key amino
acids to the intestinally digested amino
acid pool. For example, solvent-extract-
ed SBM provides 2.10 g/kg of methionine.
Expeller SBM, on the other hand, pro-
vides 2.92 g, or 39% more, of this nutri-
ent. In comparison, CM provides 2.33 g
of methionine, or 11% more methionine
for the same weight of solvent-extracted
Depending upon which nutrient is
most critical to the formulation, the
value of a lower CP, lower available RUP
ingredient might actually be greater than
Animal performance effect?
In total, seven studies conducted at the
Dairy Forage Research Center showed
that performance is numerically im-
proved when cows receive CM than when
they receive SBM. A summary of the re-
sults are presented in Table 6. There was
a consistent improvement with the more
balanced protein source.
Lending further support to production
advantages with CM are two recent me-
ta-analyses. Huhtanen et al. (2011) pub-
lished results from 122 studies in which
dietary protein was elevated using either
SBM or CM, and the higher dietary pro-
tein concentrations were achieved by
reducing the grain portion of the diets.
According to the meta-analysis, for ev-
ery 1 lb. increase in CP consumed, milk
production increased by 3.4 lb. when
the added protein was derived from CM
and by 2.1 lb. with SBM, resulting in a net
gain of 1.3 lb. with CM.
Using slightly different selection crite-
ria, Martineau et al. (2013) compared the
effects of replacing other vegetable pro-
teins in the diet with the same amount
of protein from CM. At the average inclu-
sion level (5.1 lb. per day) of CM, milk
yield increased by 3.1 lb. across the 49
studies used in the analysis.
In a continuation of this research,
Martineau et al. (2014) compared the
response in plasma amino acids to
changes in the protein source in the diet.
Essential amino acids were higher and
MUN was lower when cows received CM
compared to all other sources of protein.
What these data show is that calculat-
ing the value of proteins based on their
relative price is not accurate. The calcu-
lation can be somewhat improved by cal-
culating the value on the basis of avail-
able RUP and by comparing the supply
of the most critical amino acid. The val-
ue of improved or reduced performance
also merits consideration.
AminoDat 5.0. 2016. Evonik Industries
Brito, A.F., and G.A. Broderick. 2007.
Effects of feeding different protein supple-
ments on milk production and nutrient utili-
zation in dairy cows. J. Dairy Sci. 90:1816-
Brito, A.F., G.A. Broderick and S.M. Rey-
nal. 2007. Effects of different protein supple-
ments on omasal nutrient flow and microbial
protein synthesis in lactating dairy cows. J.
Dairy Sci. 90:1828-1841.
Broderick, G.A. 1992. Relative value of
fish meal versus solvent soybean meal for
lactating dairy cows fed alfalfa silage as sole
forage. J. Dairy Sci. 75:174-183.
Broderick, G.A. 2003. Effects of varying
dietary protein and energy levels on the
production of lactating dairy cows. J. Dairy
Broderick, G.A., and A.P. Faciola. 2014.
Effects of supplementing rumen-protected
met and lys on diets containing soybean
meal or canola meal in lactating dairy cows.
J. Dairy Sci. 97(E-Suppl. 1):751(Abstr.).
Broderick, G.A., S. Colombini, S. Costa,
M.A. Karsli and A.P. Faciola. 2016. Chemical
and ruminal in vitro evaluation of Cana-
dian canola meals produced over 4 years. J.
Dairy Sci. 99:7956-7970.
Broderick, G.A., A.P. Faciola and L.E.
Armentano. 2015. Replacing dietary soy-
bean meal with canola meal improves pro-
duction and efficiency of lactating dairy
cows. J. Dairy Sci. 98:5672-5687.
Broderick, G.A., M.J. Stevenson and R.A.
Patton. 2009. Effect of dietary protein con-
centration and degradability on response
to rumen-protected methionine in lactating
dairy cows. J. Dairy Sci. 92:2719-2728.
Chen, Z.H., G.A. Broderick, N.D. Luchini,
B.K. Sloan and E. Devillard. 2011. Effect of
feeding different sources of rumen-protected
methionine on milk production and N-utili-
zation in lactating dairy cows. J. Dairy Sci.
Colombini, S., G.A. Broderick and M.K.
Clayton. 2011. Effect of quantifying pep-
tide release on ruminal protein degradation
determined using the inhibitor in vitro sys-
tem. J. Dairy Sci. 94:1967-1977.
Contreras-Govea, F.E., S. Bertics, G.A.
Broderick, A. Faciola and L.E. Armentano.
2013. Lactation performance of cows fed
soybean meal or canola meal supplements.
J. Dairy Sci. 96(E-Suppl. 1):34(Abstr.).
Faciola, A.P., and G.A. Broderick. 2013.
Effects of replacing soybean meal with
canola meal for lactating dairy cows fed
three different ratios of alfalfa to corn silage.
J. Dairy Sci. 96:(E-Suppl. 1):452(Abstr.).
Hedqvist, H., and P. Uden. 2006. Measure-
ment of soluble protein degradation in the
rumen. Anim. Feed Sci. Technol. 126:1-21.
Huhtanen, P., M. Hetta and C. Swensson.
2011. Evaluation of canola meal as a protein
supplement for dairy cows: A review and a
meta-analysis. Can. J. Anim Sci. 91:529-543.
Martineau, R., D.R. Ouellet and H.
Lapierre. 2013. Feeding canola meal to
dairy cows: A meta-analysis on lactational
responses. J. Dairy Sci. 96:1701-1714.
Martineau R., D.R. Ouellet and H. Lapi-
erre. 2014. The effect of feeding canola meal
on concentrations of plasma amino acids. J.
Dairy Sci. 97:1603-1610.
Moore, S.A.E., and K.F. Kalscheur. 2016.
Canola meal in dairy cow diets during
early lactation increases production com-
pared with soybean meal. 99(E-Suppl.
Olmos Colmenero, J.J., and G.A. Brod-
erick. 2006a. Effect of amount and ruminal
Feedstuffs, June 5, 2017
degradability of soybean meal protein on
performance of lactating dairy cows. J. Dairy
Olmos Colmenero, J.J., and G.A. Brod-
erick. 2006b. Effect of dietary crude protein
concentration on milk production and nitro-
gen utilization in lactating dairy cows. J.
Dairy Sci. 89:1704-1712.
Paula, E.M., M.A.C. Danes, N.E. Lobos,
G.I. Zanton, G.A. Broderick and A.P. Faciola.
2015. Effects of replacing soybean meal with
canola meal or heat-treated canola meal on
performance of lactating dairy cows. J. Dairy
Sci. 98(Suppl. 2):387(Abstr.). ■