Rationally Approaching the Estimation of the Nutritive Value of Feed Ingredients and the Development of Nutritional Specifications for Aquaculture Species

Lessons from the Construction of the Asian Aquaculture Feed Formulation Database (AAFFD)

Dominique P Bureaudbureau@uoguelph.ca

Tel: +1-519-241-5533

A very large scale effort : Investment > USD 150,000 (direct) > 20 people worked on this project at the University of GuelphReviewed and compiled information from > 1000 documentsBest effort but also an imperfect one

This part of the project involved compiling or generating information on chemical and nutrient compositions and nutritive value of a large number of feed ingredients that could potentially be used in the manufacturing of aquaculture feeds

FICD

• Compiled information on about 400 generic ingredients for 239 parameters (!?)• No single study / document contained all this massive amount of information• Multiple observations for same ingredients (protein, lipid, amino acids, etc.)• Many “blank” for many/most parameters that had to be estimated

Animals Utilize NUTRIENTSnot “Ingredients”

What’s important in feed formulation?

– Individual nutrient requirements of animals (with adequate safety margins)

– Nutrient content of feed ingredients and associated variability

– Digestibility and bio-availability of nutrients

– Potential limitations (e.g. contaminants, anti-nutritional factors)

– Impacts (e.g. physical properties, waste outputs, final product quality) of the ingredients

General “mind-frame” underlying the development of the Asian Aquaculture Feed Formulation Database

10 Heads and 10 Tails: Dr. Young Cho’s Parable About

Making Sure Results are Adding Up

10 fish11 tails (?) 9 heads (?)

May be only wrong by 10% but illogical!

Nothing is lost,

nothing is

created,

everything is

transformed.

Law of Conservation of Mass

General “mind-frame” underlying the development of the Asian Aquaculture Feed Formulation Database

Ingredient PA01 PA03 PA04 PA05 PA06 PA07 PA08 PA09 PA10 PA11 PA12

Dry

Matter

Crude

Protein

Crude

Lipids

Crude

Fibre Ash NFE NDF ADF

Total

CHO Starch Sugars

% % % % % % % % % % %

Fish meal 90.8 74.2 5.0 0.5 10.0 1.2 0.0 0.0 1.7 0.0 0.0

Wheat

middlings 90.0 15.8 3.0 7.0 3.6 60.6 3.0 13.0 67.5 31.5 3.0

Canola

meal, exp. 89.9 35.2 7.5 11.9 7.0 28.4 33.3 26.0 40.3 0.9 6.0

Proximate Analysis + Carbohydrates

NRC (2011)

Crude fiber

*

*

***

*

*

*

Total Dietary Fiber

NRC (2011)

Crude fiber

*

*

***

*

*

*Digestible Not Digestible

Nutrient Contents

FAO 2.0 FEED INGREDIENT SURVEY

Variability of Lysine Concentration in Relation to Crude Protein

Content of US Soybean Meal Samples

Data courtesy of Paul Smolen and United Soybean Board

Digestibility of Nutrients

Apparent digestibility coefficients (%)

Ingredients Dry

Matter

Crude

Protein

Lipid Energy

Alfalfa meal 39 87 71 43

Blood meal

ring-dried 87 85 - 86

spray-dried 91 96 - 92

flame-dried 55 16 - 50

Brewer’s dried yeast 76 91 - 77

Corn yellow 23 95 - 39

Corn gluten feed 23 92 29

Corn gluten meal 80 96 - 83

Corn distiller dried soluble 46 85 71 51

Feather meal 77 77 - 77

Fish meal, herring 85 92 97 91

Meat and bone meal 70 85 - 80

Poultry by-products meal 76 89 - 82

Rapeseed meal 35 77 - 45

Soybean, full-fat, cook. 78 96 94 85

Soybean meal, dehulled 74 96 - 75

Wheat middlings 35 92 - 46

Whey, dehydrated 97 96 - 94

Fish protein concentrate 90 95 - 94

Soy protein concentrate 77 97 - 84

Cho and Slinger (1974); Cho and Bureau (1997)

Apparent Digestibility of Poultry By-Products Meals in Rainbow Trout

Guelph System

ADC

Protein Energy

68% 71%Cho et al. (1982)

Bureau et al. (1999) 87-91% 77-92%

74-85% 65-72%Hajen et al. (1993)

96% N/ASugiura et al. (1998)

What values do you choose / trust???

Guelph System

ADC

Protein Energy

96-99% 92-99%Spray-dried

85-88% 86-88%Ring-dried

84% 79%Steam-tube dried

Bureau et al. (1999)

82% 82%Rotoplate dried

Different drying technique

Apparent Digestibility of Different Blood Meals in Rainbow Trout

????

??

??

??

DE proximate = 1000*((.625*.46*23.6)+(.153*.622*39))/4.184 = 2508 kcal/kgDE gross energy = 4993*0.717 = 3580 kcal/kg

Importance of Being Rational and Critical in Review of Scientific LiteratureEven if data is from a reputed/famous laboratory!

a marine fish species

10 Heads and 10 Tails: Dr. Young Cho’s Parable About

Making Sure Results are Adding Up

10 fish11 tails (?) 9 heads (?)

May be only wrong by 10% but illogical!

Energy

(%)

International

Feed Number

Atlantic

Salmon

Rainbow

Trout

Atlantic

Cod

European

Sea Bass

Hybrid

Striped

Bass

Silver

Perch

Red

Drum

Channel

Catfish

Nile

Tilapia

Siberian

Sturgeon

Haddock Gilthead

Sea

Bream

Cobia Large-

mouth

Bass

Pacu Rockfish White

Shrimp

Blood meal 5-00-381 – 80–99b,c – 92f – 100i 58j – – – – 58–90s, t – – – 84–86x –

Casein 5-01-162 100a 92d – – – – – – – – – – – – – – –

Canola meal 5-06-145 – 56–75q 61e – – 58i – – – – 60r – – – – – –

Corn, grain 4-02-935 – – – – 41h – 56 j 57o 61y – – – – – 76w – 601

Corn gluten

meal

5-28-242 – – 83e 87g – 100i – – 83–89l,y – 81r 80t 94u 77v 86w 89x 872

Cottonseed

meal

5-01-621 – – – – 65–73h,z 53i 70k 80o – 72q – 39t – – – 41–55x 611

Feather meal,

hydrolyzed

5-03-795 – – – – – – – – – 80q – 64t 96u – – 73–85x –

Fishmeal, not

specified

– 99 d – – – 89–98i – – 89y 93q – 80s – 78v 75w 90–99x 872

Fish, Anchovy

meal

5-0-985 – – 86e – – – – – 92l – – – – – – 93–96x –

Fish, Herring

meal

5-02-000 – – 93e – – – – – – – 92r 94t – – – – –

Fish, Menhaden

meal

5-02-009 – – – – 96h – 60–92j,k 92o – – – – – – – – 751

Krill meal 5-16-423 – – 96e – – – – – – – – – – – – – –

Lupin meal 70a 64d 75e 87g – 51–70i – – – – – – – – – – 773

Meat and Bone

meal

5-00-388 – 68–83b – – 80h 75–81i 86k 76o – 75q – 69t 90u – – – 62–851,2

Meat meal – – – – – – – – – – – 78s – – – 90–93x 762

Peanut meal 5-03-650 – – – – – 77i – – – – – – 84u – – – 822

Poultry

byproduct meal

5-03-798 – 75–87b,c 71e 97f – – 72k – 59m 86q – 80t – 85v – – –

Poultry meal – – – – – 94i – – 79y – – 78–80s 91u – – – –

Feather meal,

hydrolyzed

5-03-795 – 76–80b,c 59e – – 100i – 67n – – – – – – – – –

Rapeseed meal 5-03-871 – 76p – – – – – – 57y – – – 83u – – – –

NRC (2011)

Apparent Digestibility Coefficient of Energy of Different Ingredients in Different Species

Being open-minded but rational and critical

Challenge:

Predicting digestible nutrient (e.g. lipids, phosphorus) contents of balanced feeds formulated to widely different digestible nutrient levels and made with a great variety of ingredients?

Number of combinations/permutations too great to study experimentally.

How can we derive the estimates we need from the literature?

It is not sufficient to know different factors have effects.

You also need to be able to quantify the combined effects of these

different factors

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50

Dietary Lipid (%)

AD

C L

ipid

(%

)

Fair amount of data but difficult to make sense of them!

Digestibility of Dietary Lipids

90 treatments from 16 studies with Atlantic salmon and rainbow trout

Modelling can be a very effective way of achieving this.

Before After

The answer is organizing the information at hand in a sensible way!

AAFFD FICD – Lipid Composition

SAFA2

-8%

SAFA*MUFA

3%

Dietary lipid

SAFA

45%

PUFA

94%

MUFA

86%

SAFA*PUFA

4%

SAFA* Water Temperature (°C)

3%

Hua and Bureau (2008)

Combined Effects of Saturated (SAFA), Monounsaturated (MUFA) and Polyunsaturated (PUFA) Fatty Acids and Water Temperature on

Digestibility of Lipids

y = -0.34x + 105

R2 = 0.87

75

80

85

90

95

100

0 10 20 30 40

Saturates (% of total fatty acids)

AD

C L

ipid

(%

)

7.5 C

15 C

y = -0.32x + 100

R2 = 0.99

75

80

85

90

95

100

0 10 20 30 40 50

Saturates (% of total fatty acids)

AD

C L

ipid

s (

%)

Apparent digestibility of lipids of diets containing different palm oil products and fed red hybrid tilapia reared at 29°C

(Data from Bahurmiz and Ng, 2007).

Apparent digestibility of lipid of diets containing combinations of fish oils and tallow and fed to rainbow trout at two different temperatures

(Bureau et al., 2008).

Contrasting Digestibility of Lipids in Different Species

y = 0.99x - 1.47

r2 = 0.98

0

5

10

15

20

25

30

0 5 10 15 20 25 30

Observed digestible lipid content (%)

Pre

dic

ted

dig

es

tib

le l

ipid

co

nte

nt

(%)

Validation of Model Prediction with Independent Dataset

Data from eight (8) fish species: Common carp (Cyprinus carpio), tilapia (Oreochromis niloticus), Atlantic cod (Gadus morhua), Murray cod (Maccullochella peelii), Australian short-finned eel (Anguilla australis), European sea bass (Dicentrarchus labrax), sharpsnout sea bream (Diplodus puntazzo), and turbot (Psetta sp.).

This part of the project involved developing nutritional specifications for a large number of commercially important aquaculture species in Asia.

This was approached in three different, complementary ways:

1) Reviewing the scientific and technical literatureCompiled information / specific data from about 1000 papers

2) Advanced nutritional modeling used to generate nutritional specification for 24 species (groups of species) and 3 to 7 life stages or weight ranges

Many “blank” for many parameters = insufficient information, need to be the focus of future efforts = Important to include now so that we start paying attention to these parameters.

ASNS

Meta-Analysis of Essential Amino Acid Requirements of Fish

Total286 studies

Relevant249

Suitable109

22 fish species

Main causes of rejection:

1) Key piece(s) of information missing in paper and preventing calculation of parameter(s) deemed important

2) Insufficient graded EAA levels (or inappropriate design for goal of meta-analysis)

3) Poor growth or feed efficiency achieved in study

109 Studies on Essential AA Requirements: Breakdown by EAA

109 Studies on Essential AA Requirements: Species represented!

22 species represented,

Too many species = dilution of research efforts

Amino Acids Atlantic Common Nile Channel Rainbow Asian European Japanese RedSalmon Carp Tilapia catfish Trout Seabass Seabass Flounder Drum Yellowtail

Arginine 1.8 1.7 1.2 1.2 1.5 1.8 1.8 2.0 1.8 1.6Histidine 0.8 0.5 1.0 0.6 0.8 NT NT NT NT NT

Isoleucine 1.1 1.0 1.0 0.8 1.1 NT NT NT NT NTLeucine 1.5 1.4 1.9 1.3 1.5 NT NT NT NT NTLysine 2.4 2.2 1.6 1.6 2.4 2.1 2.2 2.6 1.7 1.9

Methionine 0.7 0.7 0.7 0.6 0.7 0.8 NT 0.9 0.8 0.8Met+Cys 1.1 1.0 1.0 1.0 1.1 1.2 1.1 NT 1.2 1.2

Phenylalanine 0.9 1.3 1.1 0.7 0.9 NT NT NT NT NTPhe+Tyr 1.8 2.0 1.6 1.6 1.8 NT NT NT NT NT

Threonine 1.1 1.5 1.1 0.7 1.1 NT 1.2 NT 0.8 NTTryptophan 0.3 0.3 0.3 0.2 0.3 NT 0.3 NT NT NT

Valine 1.2 1.4 1.5 0.8 1.2 NT NT NT NT NT

Taurine NR NR NT NR NR R 0.2 R R R

Essential Amino Acid Requirements of Different Fish Species

Source: NRC (2011)

AQUACULTURE = Diversity of Species

>340 SPECIES

212

1542

67

3

Slide courtesy of Dr. A.J. Tacon

> 1. Tilapia> 2. Pangasius> 3. Milkfish> 4. Asian sea bass> 5. Grass Carp> 6. Common Carp> 7. Indian major carps (IMCs, 3 species)> 8. Clarias spp.> 9. Gourami> 10. Pompano

Scope : Species

> 11. Cobia> 12. Snappers> 13. Groupers> 14. Siganids - rabbitfish> 15. Snakehead> 16. L.vannamei> 17. P.monodon> 18. Macrobrachium 19. Abalone 20. Rainbow trout 21. Sturgeon 22. Pacu

What About Different Life Stages and Live Weights?

Biggest Challenge:

Developing Nutritional Specifications for Different Species, Life Stages,

Weight Ranges and Feed Types

• Formulation of feed to nutritional specifications that correspond closely to the requirements of the animal and/or production objectives without deficiency or excess

• Important step towards improving the cost-effectiveness of feeds in aquaculture

Precision Feed Formulation

Animals Utilize NUTRIENTSnot “Ingredient”, and not “Proximate Components”

What’s important in feed formulation?

– Individual nutrient requirements of animals (with adequate safety margins)

– Nutrient content of feed ingredients and associated variability

– Digestibility and bio-availability of nutrients

– Potential limitations (e.g. contaminants, anti-nutritional factors)

– Impacts (e.g. physical properties, waste outputs, final product quality) of the ingredients

This part of the project involved developing nutritional specifications for a large number of commercially important aquaculture species in Asia.

This was approached in three different, complementary ways:

1) Reviewing the scientific and technical literatureCompiled information / specific data from about 1000 papers

2) Advanced nutritional modeling used to generate nutritional specification for 24 species (groups of species) and 3 to 7 life stages or weight ranges

Many “blank” for many parameters = insufficient information, need to be the focus of future efforts = Important to include now so that we start paying attention to these parameters.

ASNS

Confidential - Please do not share or discuss without permission from Veridis

Aquatic Technologies Inc.43

Intake(100%)

FecalLosses

(10-20%)

undigested

Retained (25-55%)

Digested

Inevitable catabolism (20-40%)

Maintenance (5-15%)Endogenous gut losses

Balanced AA

Imbalanced amino acid catabolism

(10-30%)

Excess vs. potential

NH3

NH3

NH3NH3

Factorial Amino Acid Utilization Scheme

Preferentialcatabolism

NH3

Intakemg fish/day

Fecallosses

Retained

mg fish/day

Digestible Amino Acid Requirement

mg fish/day

Inevitable catabolism

mg fish/day

Maintenance

mg fish/day

Endogenous gut losses

Factorial Amino Acid Requirement Model

Efficiency of Retention

Retained methionine (g/fish) vs. methionine intake (g/fish)

RM= 0.083+ 0.427xr2 = 0.93

Inevitable catabolism = 1 – Efficiency Slope

Maintenance = intercept

Intakemg fish/day

Fecallosses

Retained

mg fish/day

Digestible Amino Acid Requirement

mg fish/day

Inevitable catabolism

mg fish/day

Maintenance

mg fish/day

Endogenous gut losses

Factorial Amino Acid Requirement Model

Factorial Model of Amino Acid Requirement Model

Absolute EAA (e.g. Met) Requirement(g per fish per day)

Divided by

Expected feed intake(g fish per day)

Equal

Optimal Dietary Concentration (%, mg/kg, kcal/kg)

How do you get this value?

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

21 42 56 77 98 119 140 161 182 203 224 245

FCR

Days

FCR (Observed) FCR (Predicted)

Observed and predicted evolution of feed conversion ratio (feed:gain) of Nile tilapia during a pilot-scale trial

0 500 1000 1500

0

1

2

3

4

5

6

0

10

20

30

40

50

60

70

80

90

Body weight (g/fish)

Fee

din

g ra

te (

% B

W/d

ay)

Fee

d R

eq

uir

em

en

t (g

/fis

h p

er

we

ek)

Predicted Feed Intake Period Feed Intake %BW/day (as-is)

Simulated feed intake of Nile tilapia of increasing weight

0 200 400 600 800 1000 1200 1400 16002,900

3,000

3,100

3,200

3,300

3,400

3,500

0

5

10

15

20

25

30

35

40

Body weight (g/fish)

Dig

est

ible

En

erg

y (k

cal/

kg)

Dig

est

ible

Pro

tein

(%

)

DP % DE kcal/kg

Predicted Optimal Digestible Protein and Digestible Energy Content of Nile Tilapia Feeds

Weight Class

Essential Amino Acids 0.2–20 g 20–500 g 500–1,500 g

% diet DM

Arg 1.91 1.77 1.62

His 0.83 0.77 0.69

Ile 1.27 1.19 0.98

Leu 2.26 2.11 1.78

Lys 2.47 2.31 1.92

Met + Cys 1.32 1.23 1.10

Phe + Tyr 2.49 2.33 1.82

Thr 1.77 1.63 1.60

Trp 0.43 0.40 0.42

Val 1.90 1.76 1.64

1

Digestible EAA requirements (% Diet DM) of rainbow trout of different weights Fed Diets with 20 MJ DE. (EAA Requirements estimated using a factorial model)

Source: NRC (2011)

Theoretical estimate of digestible P requirement of Atlantic salmon of increasing weights.

Weight Classg/fish

0.2 – 20 20 - 500 500 - 1500 1500 - 3000 3000 - 5000

Expected FCR, feed:gain*

0.7 0.8 1.0 1.2 1.6

Dig. P Requirement, Mean, %

0.74 0.55 0.44 0.35 0.25

Dig. P Requirement, Range, % **

0.91-0.64 0.64-0.48 0.48-0.39 0.39-0.30 0.30-0.20

Theoretical estimate of digestible P requirement of Atlantic salmon of increasing weights.

Estimates derived from a factorial modeling exercise (Feed with 20 MJ DE) based on the model described by Hua and Bureau (2012) and used in modeling exercises developed for the NRC (2011).

Effect of Feed Composition / Feed Grade/ Production System???

Feed is not “Feed”

Atlantic salmon (Azevedo, 1998)

Regular HND

DP, % 37 44

DE, MJ/kg 18 22

DP/DE, g/MJ 20 20

Weight gain, g/fish 33.4 33.6

Feed efficiency, G:F 1.09 1.33

FCR, F:G 0.92 0.75

0

10

20

30

40

50

Cru

de

Pro

tein

(%

)

Feeds

Protein Levels of Aquaculture Feeds Produced by a “Generalist” Aquaculture Feed Manufacturer

How you adapt the nutrient composition of feed of different chemical composition?Multiple contradictory opinions / approaches

0

0.5

1

1.5

2

2.5

10 15 20 25 30

Crude Protein (%)

Av

g D

aily

Ga

in (

g/f

ish

)

0

0.5

1

1.5

2

2.5

3

FC

R (

fee

d:g

ain

)

ADG

FCR

Daily Weight Gain and Feed Conversion Ratio of Nile Tilapia Fed Commercial Feeds with Different Nutrient Densities

Data from a commercial cage culture operation in SE Asia

- 20%

UE + ZE

Dietary DP/DE

Expected protein

retention efficiency

Actual protein gain in fish body (g/d)

BWG (g/d)

Feed intake (g/d)

DP intake

(g/d)

FE or FCR

DE intake

(kJ/d)

ME intake

(kJ/d)

HeE

RE (kJ/d)

Body lipid gain (g/d, kJ/d)

Digestible AA intake

Lipid retention efficiency

Digestible AA for deposition

Potential protein gain (g/d)

determined by AA intake

Potential protein gain (g/d)

determined by DP and DE intake

AA deposition

efficiency

Actual protein/AA

retention efficiency

Ingredient Composition DatabaseFeed Evaluation Component

A Factorial Essential Amino Acid - Bioenergetic Hybrid Model

Hua and Bureau (2012)

Lysine requirement of rainbow trout in response to different dietary protein and energy content across different body weight classes

Nutritional Specifications

• Nutritional specifications are guidelines. The are defined carefully, reviewed occasionally, and generally quite strictly followed by feed formulators to ensure consistency of nutritional quality of feeds

• Nutrient restrictions are “practical” values taking into account :

• Requirements of the animal

• Production objectives • Ex: Minimizing cost of formula while obtaining maximum performance

• Uncertainties • Ex: Uncertainties around estimate of nutritional composition,

nutritional requirements or potential losses of nutrients requiring use of certain safety margin

1- Determining nutrient requirements across life stagesEffective approach: Fine characterization of nutrient requirements

Research trials / review of literatureUse of nutritional models

2- Cost-effectively meeting nutrient requirementsEffective approach: Fine chemical characterization of ingredients

Digestibility trials, in vitro lab analysisUse nutritional models (digestible nutrients)Use additives and processing techniques

3- Verifying if predictions correspond to commercial realityEffective approach: Benchmarking / production modeling

Investment in Research & Development (R&D)Never be satisfied with status quo

Adequately and Cost-Effectively Meeting Requirements

Key Strategies:

65