Introduction – Methodology – Results – Model Evaluation – Conclusions
Characterization of Non‐nutritive Factors of Feeds for Model Development
Introduction ‐ Outline • Non nutritive aspects of feeds • Soluble AA and peptides • Key questions
Samuel Fessenden and Michael Van Amburgh Cornell University Dept. of Animal Science
• Omasal flows of nitrogen and AA • Trial design • Pools and flows of nutrients • Protozoa and bacteria growth and competition
• Characterization of microbial AA supply • Modification and evaluation of updated CNCPS v.7 1
2
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Cornell Net Carbohydrate and Protein System
Roadmap for improvement
• Digestibility = kd/(kd+kp)
• O’Connor paper outlined principle areas for improvement in protein and AA balancing:
• kd: intrinsic to the feed • kp: intrinsic to the animal
• Microbial growth rate is directly related to substrate availability (e.g., CHO kd) • Post‐rumen AA supply calculated from microbes, undegraded feed, and endogenous N secretions
3
“ 5) Further research on the amino acid content of soluble and insoluble available dietary protein escaping ruminal degradation for various natural and by‐product feedstuffs is required to increase the accuracy of predicting available amino acids supplied by specific cattle diets. 6) Additional research is necessary to determine accurate estimates of the amino acid composition of bacterial cell wall and non‐cell wall protein fractions. 7) More research needs to be conducted concerning the true digestibility of bacterial cell wall and non‐cell wall JAS, 1993 protein fractions. “ 4
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Background
Background
• Soluble AA and peptides have been shown to stimulate rumen microbial growth in vitro
• Production responses are varied • Inconsistent effects in vivo (Broderick et al., 2000) • Interaction with sugars (Penner et al., 2009)
(Cotta and Russell, 1982)
• FERMENTEN • byproduct of commercial AA production
• Need to understand mechanism to model in CNCPS • What is the fate of free AA and peptides? • Does FERMENTEN stimulate microbes, or provide RUP in the soluble phase?
• In vitro work with FERMENTEN demonstrated increased microbial protein synthesis (Lean et al., 2005) 5
6
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Our questions
Trial 1: Omasal N and AA flows
1. What do soluble AA and peptides do in the rumen? • Careful trial design to isolate variable of interest
2. Is the CNCPS structure sufficient to represent microbial growth for field application? • Microbial composition, growth, and turnover • Predicted vs. observed AA flows 7
Photo: S. Fessenden
8
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Cows and diets
Formulated soluble protein pools
• 8 cows, switchback design with 2 treatments: lb DM/d 25.4 6.8 6.8 18.0
400
% of DM 44.6 12.0 12.0 31.4
• In the protein mix, either:
N intake, g/d
TMR composition Corn silage Alfalfa haylage Corn meal Protein mix (CON or EXP)
CON EXP
300 200 100 0 CON EXP
• CON : mix of wheat midds and urea • EXP : FERMENTEN at 3% inclusion rate
Prot A1 61 43
Prot A2 171 183
Prot B1 304 310
• 20 g N shift between pools (CNCPS v. 6.5)
• 1.7 lbs per head/d at 57 lbs DMI
• 3.3% of 608 g average N intake 9
10
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Methodology
Kinetic parameters • Flow from the rumen: • Useful for evaluating and developing many types of models‐‐‐Key for empirical based systems
• Flows AND pool size (from rumen dump): • Allows for calculation of turnover and kinetics AA, % of DM
• nutrient flow / rumen pool size = fractional rate
• Useful for development and evaluation of the mechanistic elements of the model Time, h
Photos: S. Fessenden and A. Foskolos
Time, h
Time, h 11
12
N flow, g/d
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Nitrogen flows
800 700 600 500 400 300 200 100 0
Microbial N partitioning CON EXP P=0.09
Measured flows at omasum
CON
EXP
SEM
P‐value
Total microbial NAN
450
409
28
0.31
Bacteria NAN flow, g/d
378
337
23
0.22
Protozoa NAN flow, g/d
72.1
73.9
7.3
0.84
Rumen pool size
15.8 CON
17.9 EXP
1.0 SEM
0.12 P‐value
Total microbial NAN, g
340
300
21
0.21
Prot. as % of total microbial flow
Total N
NH3‐N
Item N Intake, g/d Microbial N, % NAN flow NANMN (RUN), % NAN flow Rumen N digestibility, %
NAN
CON 603 69.9 31.3 68.7
Microbial N
EXP 613 61.5 41.7 58.3
NANMN
SEM 18 3.5 3.5 3.5
P‐value 0.70 0.11 0.05 0.05
• Partition between bacteria and protozoa pools: • Do protozoa avoid passage and represent a larger portion of the pool compared with the flow? • Literature suggests 5% to 70 % selective retention
13
14
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Microbial Growth
Microbial Growth
• Calculate microbial growth: • We have flow and pool size, so: • Flow (g/h) / bact. pool size (g) = fractional growth rate
• Calculate CHO degradation • CHO dig. (g/h) / CHO pool (g)= fractional degradation rate
• Growth / degradation = Yg
Fractional Bacteria Protozoa
90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Protozoa Bacteria
0%
25%
50%
75%
Selective retention of protozoa
15
growth rate, h‐1
• CHO deg. rate: • 0.139 h‐1 • CHO allowable growth rate: • 0.069 h‐1 • Selective retention above 50% not likely
Observed Yg (g cells/g CHO)
• Russell and Wallace (1997)
100% % of rumen microbial pool
• Which retention is realistic? • Maximal yield of cells (Yg) in rumen bacteria is ~ 0.5 g of cells/ g of degraded CHO
Selective retention level 0 25% 50% 75% 0.061 0.064 0.070 0.105 0.061 0.046 0.030 0.015
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
CON EXP
0%
25%
50%
Selective Retention
75% 16
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
CNCPS v7 rumen sub‐model
Study findings
Item Observed Predicted ‐1 Fractional microbial growth rate,h 0.061 0.054 Absolute CHO digestion rate, g/h 520 485 Fractional CHO degradation rate,‐h‐1 0.133 0.124 Yield of cells/g of CHO degraded, g 0.44 0.45 Total microbial N flow, g/d 430 412 Bacteria N, g/d 357 371 Protozoa N, g/d 73 41 • Overall structure seems to predict CHO degradation and yield of cells fairly well in this study
• Feed CP degradation reduced 15% • Diet differences provided 20g more AA inflow • 212 g increase in total AA outflow
• Indicates a sparing effect on degradable protein • No indication of changes in microbial growth • Structure of rumen sub‐model seems OK for now
• Bacteria and protozoa partition needs further work
• Need more data on other products 17
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Microbial AA profile
Microbial AA profile
• Multiple time point hydrolysis
Item, % of TAA ARG HIS ILE LEU LYS MET PHE TRP THR VAL Total EAA Total AA, % of DM
• 2, 4, 6, 12, 18, 21, 24, 30, 48, 72, 120, 168 hrs • Least‐squares non‐linear regression Lys
AA, mg/g DM
AA, mg/g DM
25 20 15 10 5
Bacteria
0 0
Met
20
Protozoa
24 48 72 96 120 144 168 Hydrolysis Time (h)
Val
30 25
15
AA, mg/g DM
30
18
10 5 Bacteria
0 0
Protozoa
24 48 72 96 120 144 168 Hydrolysis Time (h)
20 15 10 5
Bacteria
0 0
Protozoa
24 48 72 96 120 144 168 Hydrolysis Time (h) 19
24 h 4.96 2.24 4.25 5.48 7.52 4.71 6.15 5.51 5.67 6.58 53.07 346.6
Bacteria Mult % ∆ 4.88 1.6 2.17 3.0 4.77 ‐12.4 5.47 0.3 7.40 1.6 4.81 ‐2.0 5.94 3.4 5.93 ‐7.7 5.70 ‐0.5 7.14 ‐8.4 51.73 2.5 339.0 2.2
24 h 5.37 2.50 4.03 6.83 8.90 3.44 6.79 4.26 4.84 4.67 51.61 295.0
Protozoa Mult 5.41 2.59 4.51 6.43 8.79 3.87 6.76 5.49 5.09 4.88 51.01 290.7
% ∆ ‐0.7 ‐3.6 ‐12.0 5.8 1.2 ‐12.6 0.4 ‐29.1 ‐5.1 ‐4.6 1.2 1.4
• Single time point hydrolysis under‐estimation of some AA (BC‐AA), resulting in over‐estimation of other AA. 20
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Model evaluation
Model evaluation‐ Lysine Original v.7
• Using updated AA profile of microbial protein • Literature dataset:
Updated v.7
• Omasal studies (n=16) with 61 treatment means • Same dataset used to evaluate v6.5 and v7.0
• Predicted N and AA flows were compared with reported values • Random effect of study included in mixed model analysis
Var. Comp.
MSPE Part. (%)
AA
R2BLUP R2MP RMSE Slope
Int.
Study
Res.
CCC
RMSPE
Um
Us
Ur
Lys old Lys new
0.92 0.58 0.94 0.60
8 27
78 81
21 19
0.36 0.76
61 23
80 6
9 21
11 73
10 20
0.69 0.79
21
22
Introduction – Methodology – Results – Model Evaluation – Conclusions
Model evaluation‐ Methionine Original v.7
Model evaluation‐ Histidine
Updated v.7
Var. Comp.
Original v.7
Updated v.7
MSPE Part. (%)
AA
R2BLUP R2MP RMSE Slope
Int.
Study
Res.
CCC
RMSPE
Um
Us
Ur
Met old Met new
0.94 0.42 3.67 0.78 0.95 0.43 10 0.67
9 8
88 89
12 11
0.6 0.4
12 21
20 66
6 11
74 23 23
Var. Comp.
MSPE Part. (%)
AA
R2BLUP R2MP RMSE Slope
Int.
Study
His old His new
0.91 0.61 4.25 0.82 0.90 0.56 7 0.74
0.8 15
74.9 25.1 0.65 70 30 0.71
Res.
CCC
RMSPE
Um
Us
Ur
25.7 9
3 8
17 34
80 58 24
Introduction – Methodology – Results – Model Evaluation – Conclusions
Introduction – Methodology – Results – Model Evaluation – Conclusions
Conclusions
Acknowledgments • Dr. Mike Van Amburgh
• Overall AA predictions were improved: Overall EAA predictions Original v.7 Updated v.7
CCC 0.66 0.69
• Van Amburgh Lab, Especially Dr. Debbie Ross
RMSPE 28.5 23.8
• Omasal flows AND rumen pool sizes helped us understand rumen effects • Techniques can be useful for investigating starch, functional CHO, yeast, monensin, enzymes etc. 25
Introduction – Methodology – Results – Model Evaluation – Conclusions
Questions
ansci.cals.cornell.edu/about‐us/history
27
• Arm & Hammer Animal Nutrition 26