Skip to main content

A whole-systems approach to optimising feed efficiency and reducing the ecological footprint of monogastrics

Final Report Summary - ECO-FCE (A whole-systems approach to optimising feed efficiency and reducing the ecological footprint of monogastrics)

Executive Summary:
It is clear that intensive production systems will play a key role in feeding a growing global population, but they must also be sustainable. This means being environmentally and animal welfare -friendly, while also allowing a reasonable return on investment for producers. Rising costs of energy and feed, combined with consumer expectation of low product prices, make this latter element a continual challenge. The pig and broiler chicken industries are key contributors to the European economy, and one of the main ways in which sustainability can be achieved is through improving feed use efficiency.

ECO-FCE employed ‛omics‘ technologies and novel systems biology research to gain a greater understanding of the biological basis of feed use efficiency in pigs and broiler chickens. It also investigated nutritional and genetic approaches to manipulate feed use efficiency and ecological footprint. The project team involved 18 partner organisations and encompassed expertise in monogastric nutrition, genetics, (meta)genomics, statistical modelling, product quality, animal production, and animal health and welfare. A major strength of the project was that it combined cutting edge science with an understanding of the realities of modern commercial farming. This was reflected in the strong contribution from industrial project partners.

A significant initial task in ECO FCE was to collate and utilise the wealth of scientific information on feed use efficiency that already existed. A key output from this element of the project was the development of a protocol which could be adopted by livestock production journals to allow automated collation of scientific data to facilitate meta-analyses. The development of nutritional tools, both in vitro and in vivo, to improve feed use efficiency and ecological footprint was also of key focus. In vitro ‘ileal digestion’ methods proved useful for the multiple screening of exogenous enzymes and additives when used in combination with different feed ingredients. This approach has potential to be used as a screening tool in future projects. Feeding trials demonstrated that non soya diets supported pig and broiler chicken performance to levels comparable to when soya based diets were offered, but beneficial effects of enzymes were variable. The use of certain plant-based additives had minimal effect on animal performance but appeared to reduce ammonia emissions in pigs, and this is an area that warrants further investigation.

Nutritional conditioning of chickens was found to increase the efficiency of absorption or metabolic utilisation of phosphorus and methionine later in life. For example, ECO-FCE results suggested a positive effect of P nutritional conditioning on performance and bone mineralisation. Specific symbiotic regimes for in-ovo use were also developed, and showed very promising results in terms of effects on performance-related parameters in broiler chickens. Trials involving early life supplementation of low birth weight pigs with L-arginine and L-carnitine did not yield the desired beneficial results, but suggested that the use of rescue decks was an effective method of promoting survival in these animals.

ECO FCE also aimed to identify key factors driving good and poor feed use efficiency in pigs and broiler chickens, when reared across different environments. Strong environmental effects were observed, and the work suggested that the physiological mechanisms underlying divergence differ for males and females. Interesting correlations with feeding behaviour, gut microbiota as well as macronutrient transport, endocrine regulation, osmohomeostasis and mitochondrial function were observed. Some novel work in this area identified the strong influence of early life microbiota inoculation on lifetime performance. At a genetic level, ECO FCE has identified biomarkers of divergent feed efficiency and has developed novel modelling methodologies to assist the breeding industries in better selecting for this trait.
Project Context and Objectives:
Project context and objectives
It is clear that improving feed use efficiency will be central to meeting future global demands for animal protein in a sustainable way. However, there is uncertaintly about how best to achieve this, and this was the impetus for the 4-year FP7 project ‛ECO-FCE‘. The project commenced in February 2013 and focused on monogastric species, namely pigs and broiler chickens, and thus complemented more ruminant-based FP7 projects such as RuminOmics.

ECO-FCE was developed to identify novel strategies to synergistically improve feed use efficiency and reduce the ecological footprint of pig and broiler chicken production systems.

The overarching objectives of ECO-FCE were:
• To improve European and global food security by optimising efficiency of feed use in the pig and broiler chicken industries
• To synergistically reduce the ecological footprint of these industries to provide sustainable eco-systems and negate climate change

These objectives were explored through five S&T work packages (Work Packages 1-5) linked to an extensive dissemination strategy (Work Package 6).

Work Package 1
It was recognised that a vast array of knowledge already existed in the peer reviewed literature with regard to the feed use efficiency of monogastrics. Therefore, the overarching aim of Work Package 1 was to harness and use this information to inform stakeholders. Whilst identifying this information, a secondary aim was also to ‘horizon scan’ the literature to identify future nutritional solutions with regard to improving monogastric feed use efficiency and reducing their ecological footprint. Within this work package scientists were also tasked with the formation of standard protocols to be adopted across experimental work within ECO FCE to capture meat quality and welfare measures. As such the three specific objectives of Work Package 1 ‘Building a foundation of knowledge, understanding and consistency’ were:
1. To summarise the current library of knowledge on monogastric feed use efficiency, organise this knowledge into a knowledge-warehouse and finally, based on this current state of knowledge, to identify knowledge gaps and summarise associations by meta-analysis
2. To identify innovative strategies through horizon Scanning
3. To set standard protocols to ensure consistency throughout ECO-FCE

Work Package 2
Optimising feeding strategies to improve feed use efficiency has great potential to reduce the ecological footprint of pig and broiler chicken production systems. Work Package 2 aimed to identify these strategies. The feeding strategies investigated included how feed was offered, as well as the composition of the diet. In pigs, improved feeding precision and different strategies of liquid feeding were evaluated. In broiler chickens, the effect of nutritional conditioning with specific nutrient deficient diets during the first week of life on nutrient utilization and performance later in life was also tested. The effect of feed composition (the nature of ingredients and supplementation with additives such as enzymes and plant extracts) on nutrient digestibility and fermentation was evaluated in vitro (using pig and poultry systems of digestion and fermentation), and the impact on feed efficiency of the most promising combinations were tested in vivo (in pigs and broilers). Particular attention was given to less-conventional ingredients, such as rapeseed meal, dried distillers grains with solubles, field beans and peas, and extruded cereals, peas and field beans, to enzymes (proteases, carbohydrases, phytase, α-galactosidase and carbohydrases), and to essential oils, spices and saponins. Use of L-carnitine and L-arginine during the early postnatal period to improve postnatal growth efficiency was also evaluated in piglets.

Work Package 3
Research conducted within this work package focused on investigating the effect of gut factors (including microbiota) on measures relating to feed use efficiency in broiler chickens and pigs. This involved performing studies with both pigs and chickens across different research sites in Europe to enhance our understanding regarding the gut microbiota composition of chickens and pigs of diverging feed efficiency and its interaction with the host animal. The aim was also to investigate the effect of the environment on feed efficiency-related gut bacterial profiles, size, structure and functionality in chickens and pigs.

Within this work package, research was also aimed at identifying optimum conditions to promote gut bacterial profiles that benefit feed use efficiency. With broiler chickens, this involved investigating two different post hatch intervention strategies to influence the intestinal microbiota in order to promote the chicken’s lifetime feed use efficiency. The first strategy was to inoculate chickens with a digesta inoculum early in life, whereas the second approach was to modify the feed intake level of chickens by comparing ad libitum feeding with restrictive feeding.
Two experiments were also conducted with pigs to investigate the timing and delivery mechanism to optimise microbial profile and ensure persistence. The objective of the first was to investigate changes in intestinal microbiota and growth in pigs following faecal microbiota transplant (FMT) in sows and/or their offspring with faecal extracts from highly feed efficient pigs. The objective of the second experiment was to determine whether FMT in pregnant sows combined with feeding inulin to progeny promoted the proliferation and persistence of an optimum microbial profile (in terms of microbial diversity, composition and predicted functionality) for improved FE throughout the productive life in pigs.

Research within Work Package 3 was also aimed at developing synbiotics for use with broiler chickens to promote optimum feed use efficiency in later life. This involved research aimed at characterizing and comparing selected lactobacilli strains isolated from different environments (milk and hen feces) regarding their persistence in the gastrointestinal tract and potential application as chicken synbiotics for administration in ovo. Additional work aimed to design and test novel synbiotics in vitro and to validate them in Cobb chickens in a commercial field study after in ovo injection on day 12 of incubation. In this study, male chickens (n=2,040) were reared for 6 wks to study physiological and performance parameters (body weight, feed conversion ratio, meat quality, gene expression in gut-associated-lymphatic tissue (GALT), intestinal histomorphology and the microbiome.

Work Package 4
The functional biodiversity of livestock species regarding their feed efficiency (FE) provides a huge potential for selective breeding. Therefore, the molecular analyses within Work Package 4 focused on the elucidation of the genetic architecture of feed efficiency related traits in chicken and pigs. Moreover, inductive and deductive strategies were applied to identify networks of genes, their molecular mechanisms of action, and underlying pathways in the context of FE. Ultimately, this strategy aims to provide results significant for both science and animal breeding.
Different methods of genome-wide association studies (GWAS) were applied to the datasets in order to identify links between genomic variation and variation in FE and related growth, carcass and meat quality traits. Therefore, resource populations of highly relevant breeds/lines in pig and poultry production were employed. These comprised about 900 broiler chickens from a line bred for improved FE, and about 1300 boars from a highly competitive sire line. Specifically, advanced methodology approaches, recently used in the field of animal genetics, like entropy analysis and the Bayesian approach, were employed to identify the most informative chromosomal regions and candidate genes as well as the interactions between most informative pairs of SNPs.

To monitor and assess individual responses and molecular alterations related to improvements in feed efficiency, holistic `omics´ techniques were applied at different levels of the genotype-phenotype axis to identify the molecular routes relevant for the efficient transition of nutrients into animal tissues. This approach was used to integrate collected data in order to identify pathways that are statistically up or down regulated, or simply over-represented, in more FE compared to less FE animals.

Ultimately, meat quality was monitored as the main output value from agri-food production impacting on human nutrition. The understanding of interactions between the molecular foundation of FE at the host side, the microbiota and the quality of products builds the basis to ensure food security and reduce the ecological footprint of monogastric species.

Work package 5
Genetic modelling of economically important traits in livestock breeding has resulted in substantial genetic progress in commercial populations. Better quantitative genetic models and tools result in the accurate estimation of the genetic merit of animals leading to cost effective improvements in traits of interest. Some key overall objectives of this work package were to develop models and tools to support breeders/farmers to detect the elite animals for feed efficiency and production traits, as well as to better understand efficiency related traits, and also to ensure the high uptake of tools through industry consultation. This research involved close co-operation between academics and breeding companies involving the assessment of several novel methodologies and approaches. This involved an evaluation of the use of longitudinal feed intake information (recorded in routine breeding programmes) in selection models, and of the use of residual feed intake as a measure of feed efficiency rather than other measures such as feed conversion ratio. Recently, large scale genomic information on animal characteristics has become available, and this requires novel genomic modelling to integrate this information into routine breeding programs. The development of integrative genomic approaches is expected to better characterise animal performance, especially for difficult and expensive traits such as feed efficiency, and this formed a significant element of work within Work Package 5.

The objective of this work package was also to develop web-based tools for use by the broader agro-food sector, such as the eECO-FCE web-tool, which allow easy access by the broader agri-food sector to findings from this project. In addition, 3-in-1 monogastric ecological calculators were developed for the broiler and pig sectors to assist the broader agri-food sector in understanding the ecological impact of management decisions relating to livestock production in these systems.

Work package 6

Promoting the use and dissemination of project results has been a key objective of the Seventh Framework Programme (FP7) and the ECO FCE project. The overall objective of ECO-FCE has been to integrate the research findings to stakeholder organizations such as policy makers, commercial industry, scientists and the general public. The ultimate goal was to provide the EU pig and poultry industries with strategies and tools to feed a growing global population in an efficient and ecologically-friendly way.

A thorough dissemination strategy was fundamental in communicating the benefits and positive impact of the uptake of ECO-FCE to stakeholders. In addition, the project had also a dissemination activity dealing with the general public: the ECO FCE exhibition. This exhibition was shown in three different science centres in Europe (Aberdeen/UK, Wels/A, Tallinn/EST). More information about this dissemination activity is described in deliverable D6.8 Report on ECO-FCE travelling exhibition.

Dissemination aims and objectives
The aims of the dissemination were to promote & disseminate ECO-FCE findings towards relevant stakeholders at a regional, national and EU level. The dissemination objectives were:
• To promote the ECO FCE project and its results as widely and as effectively as possible to all stakeholders.
• In order to do so, the project established regular communication and dialogue with relevant stakeholders and other agri-food scientists, agri-food related EU projects, and national and international organizations.
• A wide range of dissemination tools, such as a public website, flyers, posters, factsheets and publications for the public were used.
• In the context of dissemination, ECO-FCE also organized international conferences and international stakeholder events (workshops) and produced a project movie.
• A special ECO-FCE exhibition was designed and shown in different European Science Centres and Universities.
• A sustainability and IPR strategy was developed.
Project Results:
Work package 1

1. Identification of novel feed additives and ingredients to beneficially affect pig and broiler chicken feed conversion efficiency (FCE) and ecological footprint.

Innovative approaches involving novel feed supplements and feed processing technologies that could improve feed conversion efficiency (FCE) and reduce the ecological footprint of pig and poultry production were identified through an objective and systematic review of the peer-reviewed and non peer-reviewed research literature.
For pigs, diet supplementation with sunflower husks (80 g/kg), olive soap stock (62.5 g/kg), levan-type fructan (0.1 g/kg) and 0.2% resveratrol improved average daily gain (ADG) by 100, 87, 80 and 60 g/d respectively, compared to control diets, whilst average daily feed intake (ADFI) was improved by 710, 84 and 60 g/pig respectively, by sunflower husks (400 g/kg), resveratrol (0.2%) and B. Juncea canola expeller (120 g/kg). Gain:feed ratio was improved by 0.05 when brown seaweed laminarin was offered to pigs at 0.30 g/kg of diet. Additionally, pigs fed camelina meal (7.4 g/kg) as a diet supplement showed an FCE improvement of 1.4 compared to the control diet.
For poultry, diet supplementation with whole yeast product (P. guillermondii) at 0.1% of the diet, chilli powder (30 mg/kg) or leucine (0.05%) improved ADG by 13, 5.5 and 3.5 g/d, respectively, compared to control diets. Likewise, additions of M. stenopetala at 140 mg/kg of diet, or wet fermented soybean meal (at 6% of diet) or pennywort (at 0.5% of diet) improved ADFI by 14.2 6.5 and 5.4 g/bird, respectively, compared to the control diet. Furthermore, feeding broilers ghrelin (50 mg/kg), P. guilliermondii (0.2%), and M. stenopetala (80 mg/kg) improved FCR by 0.28 0.24 and 0.2 respectively, compared to the control diet.

In summary, eighteen diverse feed supplements and feed additives that were found to have a positive effect on pig performance included: Bacillus subtilis, Brassica juncea canola expeller cake, Brassica juncea canola meal, Brassica juncea light particle fraction, brown seaweed laminarin, Camelina sativa, Erythrina glauca, Faba bean, Helianthus, levan-type fructan, Spirulina maxima, olive soap stock, Pisum sativa, Resveratrol, Stevia rebaudiana, Smectite and Trichothecenes. A further nineteen supplements and additives found to have positive effects on poultry performance including: bamboo vinegar, Bazhen, bromelain, Helianthus, leucine, chilli meal, curcumin, cypermethrin, Forsythia suspensa, garlic powder, lemon peel, Linum usitatissimum, Moringa stenopetala leaves, oregano, Penywort, Pichia gillermondii, Prunus mume, rosemary and Ulva lactuca. The selective in ovo administration of ghrelin as an appetite stimulant in hatchlings was also noted as a treatment option.

2. Establishment of a mega eECO FCE warehouse of information to conduct mini meta analysis.

Following a large systematic review and extraction of information from over 2500 peer review papers an eECO FCE warehouse of knowledge was built. Within this warehouse, the pig database contains the results from 810 papers, majority of which were published between 2008 and 2013. A total of 8237 ‘lines’ of information is recorded. Therefore 8233 ‘treatments’ are included in this database. The poultry (broiler) database contains the results of approximately 1700 papers, the majority of which were also published between 2008 and 2013. A total of 17,074 lines of information was recorded, reflecting results of 17,074 treatments.
For both pigs and poultry the vast majority of information is recorded against ‘feed’ (approximately 85%), with the remaining information reflecting information taken from ‘gut’ and ‘genetics’ categorised papers. However, it should be noted that a large proportion of the feed papers also contained ‘breed’ information and some ‘gut’ information.

The following lists the information housed in the excel-based eECO FCE warehouses for pigs and broiler chickens.
Reference information
First author, Year, Paper title, Journal title, Journal volume, Journal page(s)
Materials and Methods
Species, Breed, Age start (weeks), Age end (weeks), Gender

Number of animals used per treatment, Treatment name (each treatment as a separate row of data), Is this the control treatment?

Type of study
Classical quantitative study based on pedigreed populations, Genomic studies using genetic markers, Comparison of different genotypes

Flooring, Space allowance (number), Space allowance (units), Health status

Feed system
Frequency of feeding, Feed form, If other, please state, Feed form (further specification), If other, please state

Feed composition : Ingredients
Ingredient 1
Name, Level, Units
Ingredient 2........... Ingredient 17
Additive type 1 etc
Insert the feed additive ‘type’ noted within ‘ingredients’. Additive 1 should correspond with the first feed additive listed etc

Nutrient composition
Is nutrient composition based on chemical analysis or formulation analysis?
Nutrient 1 etc
Nutrient measured (name), Level (number), Units

Sample preparation
Storage details, Method used for microbe/bacteria determination,

Challenge study?

Production measures
Are the statistics robust?
Start weight in kg (mean per treatment)
Start weight (SD per treatment)
End weight in kg (mean per treatment)
End weight (SD per treatment)
Hot or cold end carcass weight? (Hot/Cold)
P-value for difference in end weight between this treatment and control
Carcass Feed Conversion Ratio (mean per treatment)
Carcass Feed Conversion Ratio (SD per treatment)
P-value for difference in carcass feed conversion ratio between this treatment and control
Carcass weight in kg (mean per treatment)
Carcass weight (SD per treatment)
P-value for difference in carcass weight between this treatment and control
Average Daily Gain in grams (or equivalent) per treatment
Average Daily Gain (or equivalent) (SD per treatment)
P-value difference in average daily gain between this treatment and control
Average Daily Feed Intake in grams (or equivalent) per treatment
Average Daily Feed Intake (or equivalent) (SD per treatment)
P-value for difference in average daily feed intake between this treatment and control
Feed Conversion Ratio (or equivalent: i.e. ADFI/ADG) (mean per treatment)
Feed Conversion Ratio (SD per treatment)
P-value for difference in feed conversion ratio between this treatment and control
Residual Feed Intake (mean per treatment)
Residual Feed Intake (SD per treatment)
P-value for difference in residual feed intake between this treatment and control
Gain to Feed (i.e. ADG/ADFI) (mean per treatment)
Gain to Feed (i.e. ADG/ADFI) (SD per treatment)
P-value for difference in gain to feed between this treatment and control
Back fat depth in mm (mean per treatment)
Back fat depth (SD per treatment)
Back fat position measured
P-value for difference in back fat depth between this treatment and control

N excretion (mean per treatment)
N excretion (SD per treatment)
P-value for difference in N excretion between this treatment and control
P excretion (mean per treatment)
P excretion (SD per treatment)
P-value for difference in P excretion between this treatment and control

CO2 emission (mean per treatment)
CO2 emission (SD per treatment)
P-value for difference in CO2 emission between this treatment and control
CH4 emission (mean per treatment)
CH4 emission (SD per treatment)
P-value for difference in CH4 emission between this treatment and control
NH3 emission (mean per treatment)
NH3 emission (SD per treatment)
P-value for difference in NH3 emission between this treatment and control
N2O emission (mean per treatment)
N2O emission (SD per treatment)
P-value for difference in N2O emission between this treatment and control

Meat quality
Drip loss measured?
Ultimate pH measured?
Colour measured?
Shear force/instrumental tenderness measured?
Eating quality measured?

Animal welfare
Does the paper mention health/welfare
If yes, is health/welfare re:

Effect of treatment on morphology
Villus height in μm (mean per treatment)
Villus height (SD per treatment)
Crypt depth in μm (mean per treatment)
Crypt depth (SD per treatment)

Gut measures
Effect of treatment on digestibility
Does the paper report ileal digestibility of nutrients?
Total tract digestion of energy in % (mean per treatment)
Total tract digestion of energy (SD per treatment)
Total tract digestion of crude protein in % (mean per treatment)
Total tract digestion of crude protein (SD per treatment)
Total tract digestion of dry matter in % (mean per treatment)
Total tract digestion of dry matter (SD per treatment)
Effect of treatment on Barrier function/permeability
Short circuit current in μEq/(cm2*h) (mean per treatment)
Short circuit current (SD per treatment)
Tissue conductance in mS/cm2 (mean per treatment)
Tissue conductance (SD per treatment)
Effect of treatment on enzyme activity
Has pancreatic enzyme activity been reported
If yes, state kind of enzyme
Pancreatic in nmol/min/mg protein (mean per treatment)
Pancreatic (SD per treatment)
Has brush border enzyme activity been reported
If yes, state kind of enzyme
Brush border in nmol/min/mg protein (mean per treatment)
Brush border (SD per treatment)
Effect of treatment on gut length
Small intestine in cm (mean per treatment)
Small intestine (SD per treatment)
Large intestine in cm (mean per treatment)
Large intestine (SD per treatment)
Effect of treatment on gut mucosa inflammation score
What score used? (name)
Median score (per treatment)
Score range (minimum per treatment)
Score range (maximum per treatment)
Effect of treatment on faeces score
What score used? (name)
Median score (per treatment)
Score range (minimum per treatment)
Score range (maximum per treatment)
Effect of treatment on VFA
Large intestine in μmol/g or ml (mean per treatment
Large intestine (SD per treatment
Effect of treatment on gut mucosal immune cells/cytokines
In small intestine?
If yes, state what kind of cells
In large intestine?
If yes, state what kind of cells
Effect of treatment on microbe/bacteria composition
Have microbe/bacteria compositions been reported in stomach
If yes, what species are reported?
Have microbe/bacteria compositions been reported in small intestine?
If yes, what species are reported?
Have microbe/bacteria compositions been reported in large intestine?
If yes, what species are reported?
Effect of treatment on passage rate
Ileal (mean per treatment)
Ileal (units)
Ileal (SD per treatment)
Total tract (mean per treatment)
Total tract (units)
Total tract (SD per treatment)
Has gene expression of gut mucosa/pancreas related to digestion, absorption (transporter), barrier function, immune response reported in:
Small intestine?
Large intestine?
Carcass Feed Conversion Ratio (number)
Average Daily Gain (or equivalent) (number)
Average Daily Feed Intake (or equivalent) (number)
Feed Conversion Ratio (or equivalent: i.e. ADFI/ADG) (number)
Residual Feed Intake (number)
Gain to Feed (i.e. ADG/ADFI) (number)
Back fat depth (number)
Other genetic measures reported
Does the paper report metabolites?
Does the paper report proteins?
Does the paper report SNPs?
Does the paper report breeding value accuracies?
Does the paper report transcripts?
Does the paper report QTLs?

Genetic correlations between traits
For each of the traits reported in results:
Trait 1 (select trait)
Trait 2 (select trait)
Genetic correlation (number)
Associated standard deviation
Associated P-value

The key conclusions from this element of research were:
1) The scale of this exercise was too big, but it is suggested that a more focused approach and the use of fewer individuals for information extraction could be successful.
2) It is surprising that automation with regard to the collation of information on animal performance is not present within key scientific journals. This represents a gap which should be addressed to facilitate future meta-analysis exercises. This project has developed knowledge and experience on how such automation should or could be achieved to capture the vast amount of information being generated daily in this area.
3) There is a lack of information on the interaction between chicken genotype and diet on the parameters of interest, but this is perhaps of lower priority since the two main broiler genetics companies involved work actively in this area and have sophisticated nutritional protocols.
4) There is also a lack of information on the interaction between pig genotype and diet. The data captured in this exercise suggest that many interactions occur, in particular between protein content of the diet and both animal gut structure, animal performance and diet digestibility. This exercise has highlighted a gap in knowledge it this area which will limit the advancement of ‘precision nutrition’.
5) In agreement with previous research, phytase addition and crude protein level are the key drivers of phosphorus and nitrogen excretion respectively. It is appears that their effects are largely consistent across genotypes when comparing diverse treatments
6) There is a lack of knowledge in the area of greenhouse gas emissions from pigs and poultry. Whilst these animals do not emit as much GHG’s as ruminants, they do outnumber ruminants at a global level. Therefore more knowledge is required to understand their contribution to global emissions and the role they can play in reducing them.

Legacy from Work Package 1 research
Many lessons have been learnt on how best to extract information from the literature for the purpose of ‘mega’ meta-analysis, and the level of information which is manageable. The ‘key words’ established and the templates generated to capture the information in this exercise form key bases to build an automated system for information capture. Initial discussions have occurred with a high impact animal science journal with regard to the development of an automated information capture system which would facilitate subsequent automated meta-analysis of data. Initial support has been given and discussions will continue to advance this concept post project since it has the potential to significantly advance the ability of the animal science community to understand and fully utilise published information.

Work Package 2
Precision feeding of pigs
In an initial trial, the growth (protein deposition) and feed intake curves during the growing-finishing phases of the progeny from Hermitage Sows and MaxGro Hermitage boars were characterised. The data obtained were used to estimate nutrient requirements and their evolution with age for each sex.
This was followed by two simultaneous precision feeding studies with entire males or females, respectively, that evaluated the effects of improved precision feeding on feed efficiency and environmental impact (vs. conventional feeding). In these studies, the use of conventional feeding (with either two or five feeding phases), in which the population average requirement is used as a feeding guideline, was compared to improved precision feeding with up to 9 feeding phases. In the latter approach, three different guidelines were used, as pigs were split into 3 homogeneous groups (heavy, intermediate and light animals) and each group was fed according to its own average requirement. In the case of the conventional five phase feeding regime, it was also applied using homogeneous pens with heavy, intermediate or light animals, to study any possible effects of pen homogeneity not related to feeding strategy.

It was observed that improving feeding precision in growing-finishing pigs (by using more homogeneous groups of animals and up to 9 feed compositions) resulted in limited advantages relative to conventional phase (2 feed changes) and multiphase feeding (5 feed changes). However, during the finishing phase and particularly in males, an improvement in feed efficiency was observed for the treatments with improved precision. This suggests precision feeding may be a valid way of improving efficiency, although further improvements in precision may be required.

Liquid feeding of pigs
Liquid feeding (and prolonged soaking) may facilitate the action of both endogenous and exogenous enzymes prior to feeding and improve nutrient digestibility and feed efficiency. Different liquid feeding approaches (with or without enzymes and or prolonged soaking of the diets) were evaluated in 4 trials.
In the first trial, wheat-barley based diets were supplemented with exogenous β-glucanase and xylanase under dry and liquid feeding regimes, and in the case of liquid feeding, feed was offered with and without prolonged soaking prior to administration. Prolonged soaking of the diets supplemented with enzymes appeared to be the most effective combination to improve performance and nutrient utilization, and particularly, resulted in improved phosphorous digestibility.
In the second trial, the addition of exogenous β-glucanase, xylanase, and phytase with and without prolonged soaking prior to administration was tested in maize-wheat-barley-maize DDGS diets. The major impact of using enzymes was an improvement of phosphorus utilization. Soaking of feed also improved mineral utilization (P in particular).
In the third trial, liquid feeding in combination or not with exogenous β-glucanase and xylanase and with and without continuous fermentation (of cereals only) prior to administration was tested. Fermenting the cereal fraction of the diet improved growth, final live weight and carcass weight, but reduced lean meat percentage in the carcass of pigs. Supplementing enzymes to fresh or fermented diets improved the feed efficiency of pigs.
Finally, the fourth trial tested liquid feeding in combination or not with exogenous β-glucanase and xylanase and with and without prolonged soaking (of the cereal fraction only). Soaking the cereal fraction of the diet prior to feeding showed potential to increase growth rate in liquid fed pigs while enzyme supplementation had no positive effect on growth performance and actually increased carcass fatness.

Nutritional conditioning in broiler chickens
Nutritional conditioning (i.e. feeding of a nutrient deficient diet early in life to improve the efficiency of how the animal will use it in later stages) was also tested in two trials with broiler chickens: one for phosphorous and one for methionine. A tendency for a positive effect of early P conditioning on weight gain and feed to gain ratio was observed in the phosphorous trial. Early P conditioning improved bone mineralisation (total tibia ash), although no effect on AID of P could be detected. In P conditioned birds, an increase in the expression levels of Type IIb sodium-phosphate cotransporter was observed in the duodenum. Early Met conditioning tended to improve feed to gain ratio birds fed a Met limiting diet later in life.

In vitro digestibility and fermentability
Several combinations of feed ingredients with different exogenous enzymes were tested with an in vitro digestibility method to simulate pig and poultry digestibility. The estimated efficacy of the different exogenous enzymes tested in vitro varied depending on the type of feed ingredient used as substrate. Substantial differences were also observed between the two in vitro simulations used (for pigs and for poultry, respectively). β-glucanase improved NSP digestibility in barley (in the poultry model). The combination β-glucanase+β-xylanase improved NSP digestibility in wheat (in the pig model), DM, CP and starch digestibility in wheat (poultry model), and NSP digestibility in wheat and wheat DDGS (poultry model). α-galactosidase (at 100 times the recommended dose) improved NSP digestibility in soybean meal (in the poultry model) and DM and starch digestibility in peas (in the pig model). Protease improved DM and CP digestibility in wheat and soybean meal, and DM in barley and maize DDGS (in the poultry model). 6-phytase improved DM and NSP digestibility in maize DDGS (in the pig and poultry models). The in vitro fermentation studies, using the undigested residues from in vitro digestion generally demonstrated a reduced fermentability when feed ingredients were supplemented with exogenous enzymes. Ingredient nature had a clear effect on the microbiological profile of the fermentation residues, whereas no effect was observed for most enzymes, except for protease that resulted in an increased proportion of Entererobacteriaceae. The inclusion of plant extracts resulted in increased gas production for both, the pig and poultry models. The results obtained with these in vitro studies were considered for the design of the in vivo trials conducted with pigs and with poultry.

In vivo screening of selected enzymes in pigs
A first series of 4 trials was conducted in pigs to evaluate the efficacy of different enzyme combinations.
The first trial focused on the use of enzymes in diets where the majority of protein was delivered from non soya ingredients. Ten diets were used to examine the effects of added protease and a xylanase/ß-glucanase mixture. Diets were rape seed meal (RSM) and DDGS based except for a soya bean positive control. Phytase was included in all diets (with a reduction in P and Ca levels), available amino acids (AA) were reduced by either 3 or 8%, and energy (NE) was reduced by either 5 or 10% to test the effects of the protease and xylanase/ß-glucanase, respectively. The diets were offered to 800 growing/finishing pigs, (pens of 10 from 17 to 103kg). There were no strong effects of enzyme addition to the RSM/DDGS based pig diets. Reducing AA and energy in the growing pig diets had no effect on feed intake (average 1182 g/day) or growth rate (average 656 g/day) (P>0.05) but FCR deteriorated as the level of AA and energy reduction increased (P<0.01) (from 1.72 for the RSM/DDGS control to 1.95 when AAs were reduced by 8% and energy by 10%). Overall the impact of enzymes in these diets, even when the concentration of key components (amino acids and energy) was reduced, was insignificant. Overall, diets containing high levels of RSM and wheat DDGS supported the performance of growing and finishing pigs to levels comparable to when SBM-based diets were offered. However, enzyme addition to these diets did not markedly change pig performance.
The second trial looked at the effect of supplementing phytase, carbohydrase and protease, alone or in combination, to wheat DDGS- and rapeseed meal-based diets. A Ca and P sparing effect of phytase was observed using diets with reduced total P and Ca. No effect of carbohydrase, alone or in combination with protease and/or phytase was observed on FCR. Protease improved FCR in male but not in female pigs.
The third trial did not reveal significant effects of the addition of alfa-galactosidase-xylanase on the performance of grower-finisher pigs fed soybean meal containing diets. However, the addition of alfa-galactosidase to rapeseed meal containing diets significantly improved weight gain and feed efficiency in grower-finisher pigs.
The fourth trial tested the efficacy alfa-galactosidase-xylanase on the performance of grower-finisher pigs fed SBM-RSM (soyabean meal-rapeseed meal) containing diets, with or without a reduction in energy content. Although no effects on performance were observed, it appeared that the enzyme may have increased energy availability during the finisher phase, as suggested by a reduced feed intake (when added onto the low energy diet), and an increased carcass fat content (when added onto the positive control diet).

A second set of trials was initially designed to test the ‘optimum strategy’ across sites, but due to the inconsistent nature of results from the first set of trials these later trials (fifth to eighth) continued to test combinations of enzymes and phytogenic feed additives to identify an optimum strategy.
The fifth trial compared five wheat DDGS/RSM based diets for pigs. The control diet supplied adequate nutrients, and then this diet was lowered by 8% in net energy and 10% in amino acid content to create a negative control. To this negative control, either a combination of carbohydrases and protease were added, a phytogenic feed additive complex or both. A total of 200 animals were penned in groups of 4 between 39 and 107kg. There was no significant effect (P>0.05) of diet on animal weight, live weight gain, feed intake or feed use efficiency which averaged 844 g/day, 2404 g/day and 2.85 respectively. Phytase was included across all diets in this work as standard. It is possible that the impact of phyase masked any other positive effects of the additional enzymes and additives.
The sixth trial evaluated the effect of supplementing α-galactosidase and protease, alone or in combination, to field bean-barley based diets for pigs. Protease supplementation improved feed efficiency in the field bean based diets whereas alfa-galactosidase had no effect on pig growth or feed efficiency.
The seventh trial tested the efficacy of α-galactosidase and xylanase in SBM-RSM based diets formulated for pigs with or without a 3% overvalue for net energy and amino acids. It appeared that the enzymes might have some effect on nutrient availability, as suggested by a reduction in feed intake, weight gain and feed efficiency during the grower-2 phase and an increase in weight gain during the finisher phase.
Finally, the eighth trial tested the efficacy of α-galactosidase and xylanase in weaned pigs fed SBM based diets. Improved performance was observed during the pre-starter phase suggesting that the benefits of enzymes may be more evident in young animals.

Test of selected plant substances on feed efficiency and GHG emissions in pigs
In a series of trials, essential oils, spices and saponin sources were tested for their effect on feed efficiency and ammonia emissions.
Different active plant principles (essential oils, spices, saponins) and their combinations (essential oils + saponins; essential oils + saponins + spices) were evaluated for their potential to improve the performance and ammonia emissions in growing pigs. Spices alone or in combination with essential oils had positive effects on the growth performance. The most pronounced effects of the spices was on feed intake, which was 5% higher than in the control treatment. Best effects on feed efficiency were obtained for the combinations with saponins. Essential oils and saponins gave also the most pronounced reduction in ammonia emissions, while results were not conclusive for effects of spices on ammonia emissions.

A standardized mixture of saponins and spices alone or in combination with different essential oils was also tested for their potential to improve performance and reduce ammonia emissions in finishing pigs. Clear differences between essential oils when applied on top of a combination of saponins and spices in respect to growth performance of the pigs were shown. Saponins and spices alone did not result in an improved feed efficiency, while the addition of the essential oils led up to 2.3% improvement. On the other hand, there were no additional benefit of the essential oils compared with the addition of spices and saponins alone. The mode of action of the tested active plant principles seemed to be clearly different, and combination of different actives might not always give the best response.

The effects of different saponin sources on growth performance and ammonia emissions in finishing pigs were also investigated. There was a clear reduction of ammonia emissions achieved by the dietary application of the saponin sources. Saponins led to slightly reduced feed intake, which was either reflected in numerically lower growth, or improved feed efficiency. It can be concluded that different saponins show different effects on performance and also different potential for ammonia emissions. It also seems that there is a clear dose dependency of effects.

The effects of a standardized mixture of essential oils and saponins were tested in growing-finishing pigs. The additives were either applied during the whole fattening period or during the finishing phase only. The results were compared with an untreated control. Performance and emission data were evaluated. The experiment clearly showed the potential of essential oils and saponins to improve the performance of fattening pigs. Effects on feed efficiency were numerically better when essential oils and saponins were applied over the whole fattening period, while ammonia emissions were already substantially decreased when the plant substances were applied only during the finisher period.

The effects of different essential oils were investigated on growth performance in weaned piglets. Data were compared with a control treatment (no additives). Both of the essential oils tested tended to improve the bodyweight homogeneity of the pigs. One essential oil had beneficial effects on feed intake, most probably by the appetizing effects, while the other essential oil improved feed efficiency without increasing feed intake. The results are relevant to select the most suitable substances for pigs of different ages.

An essential oil alone or in combination with spices as well as the application of short and medium chain fatty acids on top of the treatments was evaluated in weaned piglets. The results showed again an increase in feed intake by the dietary application of spices. In contrast, short and medium chain fatty acids had rather a negative effect on feed intake, but improved feed efficiency and daily gain. The addition of spices on top of essential oils and short and medium chain fatty acids did not give conclusive results.

Different essential oils were evaluated for their effects in weaned piglets. The results showed some benefit of the combination of essential oils over the application of an essential oil alone in respect to growth performance of piglets. However, further experiments are needed to obtain more results and conclude on the best composition of essential oils.

Different essential oils were evaluated for their effects in growing pigs. Positive results of the two essential oils tested with respect to feed efficiency and ammonia emissions were shown. Reductions in emissions achieved were up to 16% and thus, were almost comparable with effects obtained with saponins. Both essential oils gave similar effects.

The dose-response effects of a saponin source on growth performance and ammonia emissions from finishing pigs were assessed. The results clearly showed the potential of saponins to reduce ammonia emissions in finishing pigs. Feed efficiency could also be improved by the application of the saponins. There was no clear dose response and the low dosage was effective.

In vivo screening of selected enzymes in poultry
The first trial focused on the use of enzymes in diets where the majority of protein was delivered from non soya ingredients. Ten diets were used to examine the effects of added phytase and protease. Diets were RSM and DDGS based except for a soya bean positive control. P and Ca levels were reduced to investigate the effect of phytase, and amino acid levels were reduced by 1.5 or 3% to test the effects of protease. The diets were offered to 600 broilers (0-35d in pens of 10). There was no significant effect on feed intake or weigh gain (average 90.3g/d (SEM 2.50) and 64.3g/d (SEM 2.12) respectively, between D 0 and D 35). When a 3% reduction in AA was adopted, FCR was poorer in the starter period (1.15 SEM 0.024) and overall between D 0 and D 35 (1.42 SEM 0.033) (P<0.05) compared with when a 1.5% reduction was adopted (FCR 1.10 and 1.37 respectively). When phytase was included FCR was improved (P<0.05) in the grower and finisher period. However there were no other significant effects of enzyme addition on bird performance. Overall the impact of enzymes in these diets, even when the concentration of key components (amino acids and energy) was reduced, was insignificant. Overall, diets containing high levels of RSM (rapeseed meal) and wheat DDGS supported the performance of broiler chickens to levels comparable to when SBM (soyabean meal) -based diets were offered. However, enzyme addition to these diets did not markedly change broiler chicken performance.
The second trial was performed with 8 dietary treatments consisting of 2 control diets based on either wheat or barley, that had been extruded or not, and with or without supplementation with carbohydrases. Barley extrusion had a negative effect on performance, which was associated with increased intestinal viscosity and reduced nutrient digestibility. Wheat extrusion increased energy digestibility but had no effect on performance. Enzyme supplementation improved nutrient digestibility, but resulted in a decrease in feed intake for both cereals, and reduced growth for wheat. Positive effects of enzymes on litter quality and severity of pododermatitis were also observed for barley.
The third trial tested 3 dietary treatments consisting of a control diet and the same diet supplemented with protease at two levels of energy and amino acids reduction. The birds receiving protease supplemented diets with reduced contents of energy and amino acids (1.5 or 3% reductions) showed significantly poorer bodyweights than the control throughout the first four weeks of the trial, although this difference was no longer significant at 37 days. Protease diets also resulted in numerically reduced cumulative feed intake. The FCR of these birds were numerically although not significantly poorer than controls, and the FCR deteriorated as the sparing effect increased from 1.5 to 3% of the diet. Based on these results it would not appear cost effective to include protease into the diet of broiler chickens although there is a potential for further study of this enzyme.
As for the pigs, the second set of trials, initially designed to test the ‘optimum strategy’ across sites, had inconsistent results, so the second set of trials (fourth to seventh) continued to test combinations of enzymes and phytogenic feed additives to identify an optimum strategy.
In the fourth trial, four dietary treatments were formulated based on soyabean meal. A positive control was offered which met or exceeded the nutrient requirements of the birds, a negative control then reduced the level of energy and amino acids in the diet by 5%. Proteases was then added to both the positive and negative controls. There were some interactive effects observed between nutrient specification and protease. Liveweight at 7d and 14d and LWG for the starter period (0-14d) of birds offered the standard positive control diet was lower (P<0.05) than those offered the 5% sparing negative control diet without protease, but there was no difference when protease was added. Overall there were no significant effects of nutrient specification on the grower, finisher or overall periods. The addition of protease improved liveweight at 7d, 14d, 21 and 28d (P<0.05) but not at 35d. Starter LWG was also improved (520.4 vs. 468.5g SEM=4.75 P<0.001) but again, the effect was superseded by the interactive effect. Protease addition improved feed efficiency (1.22 vs. 1.33 SEM=0.024 P=0.002) in the starter period but had no effect on any other parameter in the grower, finisher or overall periods. In conclusion, protease addition impacted broiler chicken performance and could be useful when improving chicken performance.
The fifth trial studied the effects of extrusion of peas or faba beans, with or without addition of exogenous enzymes (combination of xylanase, β-glucanase, α-amylase, α-galactosidase and protease). Extrusion of peas or Faba beans had little effects on viscosity and performance. Enzyme addition had no effect on performance but appeared to differentially modulate microflora depending on the legume source.
The sixth trial tested whether increasing levels of insoluble fibre (through soy hull inclusion) would benefit broiler performance in a cost effective manner. No performance benefit from supplementing soy hulls to broiler diets was observed, and from 28 days onwards, the bodyweight and FCR of the soy hull treatments was significantly poorer than the control. However, P in excreta was reduced with soy hulls, and this deserves further work.
Finally, the seventh trial aimed to determine whether double dosing an NSP enzyme would allow to reduce crude protein and amino acids by 3% without affecting broiler performance. Double dosing NSP enzymes did not compensate the lower specifications in the diets, however, further work may be needed as the recovered xylanase activity of the diets was very low.

Test of selected plant substances on feed efficiency and GHG emissions in poultry
In a series of trials different combinations of essential oils, spices and saponin sources were tested for their effect on feed efficiency and ammonia emissions.
Five saponins sources were tested on the feed efficiency and growth performance in Cobb 500 broilers in comparison with a control treatment (no addition). The diet were based on corn, wheat, soybean and rapeseed, and saponins were applied at the same dietary concentrations. Dietary saponin application improved the feed efficiency up to 2.3%. Some saponins led to numerical reduction in feed intake, which resulted also in slightly lower weight gain. It is concluded that different saponins sources exhibit different effects on performance of broiler chickens even when applied at the same concentration in feed.

A standardized combination of spices was tested at two dosages and in combination with three saponins sources, respectively. Saponins were selected based on results of the study above. All dietary treatments, except the high dose of the spices improved the gaily gain of the chickens. The combination of saponins and spices resulted in numerically better performance than application of spices alone. There were also slight differences in performance between the different saponin sources and results confirmed previous findings.

A trial aimed to evaluate the efficiency of a standardized mixture of essential oils in diets with different nutrient concentrations and under the simultaneous application of feed enzymes, i.e. phytase and NSP enzymes. The down-spec in dietary nutrients was according to the matrix values of the products tested. The down-spec in nutrients was not reflected in the performance of the chickens. The absence of any effects might be explained by higher nutrient contents in the reduced diets than expected. The application of the enzymes and the essential oils alone did not give any benefit, while the combination of phytase, NSP enzymes and essential oils improved performance to the level of the positive control (normal nutrients).

The effects of two different essential oils in combination with a defined saponins source on the growth performance in broiler chickens were compared. The objective was to optimize the composition of the essential oils used in combination with saponins. There was, however, no positive effect of the dietary applications and it concluded that the combination of the essential oils is not suitable to improve performance in broiler chicken. A revision of the formulation is necessary.

The potential of the addition of two antioxidant plant extracts on top of a mixture of essential oils and saponins was assessed. The effects of the plant extracts were evaluated as dose-response using three application levels. The target was to improve the efficiency of essential oils and saponins by herbal antioxidants and to obtain already indications for best strategy. The results of the study did not show any benefit of the supplementation of the diets with the two selected extracts. The application of saponins and essential oils alone led to slightly improved daily gain and feed efficiency.

The effects of different saponins and combinations thereof on performance and ammonia emissions was evaluated. The experiment was performed with diets based on corn-soybean meal. All treatments reduced the emissions of ammonia between 15-22% compared with the control. Daily gain and feed efficiency confirmed the potential of the previous trials, where two saponins sources gave better results than others.

Conventional and organic essential oils were evaluated for beneficial effects on growth performance in broiler chicken. The essential oils were applied in combination with saponins. The diets were based on wheat, corn and soybean. The results indicated slightly better results with the conventional oils. A final explanation cannot be given and further studies are necessary to investigate the reason for the differences.

The potential of an addition of short and medium chain fatty acids to a mixture of essential oils and saponins for improvement of growth performance of broiler chickens was evaluated. Treatments were applied during grower and/or finisher period. The diets were based on corn and soybean meal. The dietary applications improved the daily gain without affecting the feed intake. Feed efficiency was therefore improved. There was no benefit of the addition of short and medium chain fatty acids over the application of essential oils and saponins alone. Further studies are needed to evaluate the potential of such feed additive combinations under challenge conditions and on health criteria.

The target of this trial was to optimize the composition of the essential oils and saponins. Two essential oils were combined with one or two saponin sources. The diets were based on wheat, corn and soybean meal. Numerically highest bodyweight gain was achieved when a mixture of saponins was applied. However, the strongest reduction in ammonia emissions was achieved when essential oil mix 1 was combined with a single saponins source.

Essential oils and saponins were combined with two saponins sources applied in 3 dosages in order to optimize the formulation for best effects. The diets were based on wheat, corn and soybean meal. The positive effects of the selected saponins source on daily gain and feed efficiency were confirmed. The application of spices did not give any performance benefit.

Peri-natal nutritional effects on lifetime performance
Three trials were conducted. The first trial had the objective of comparing two dietary supplements [L-carnitine (CAR; 400 mg/d) and L-arginine (ARG; 1.08 g/kg body weight (BW))] to a placebo (CON) offered in the early postnatal period on growth performance, organ and semitendinosus muscle (STM) development of low birth weight (BtW) pigs. Piglets supplemented with CAR showed minor improvements with greater numerical periodic growth performance. Minor indications of altered expression of a gene related to protein synthesis at the molecular level was observed in pigs fed ARG supplemented milk replacer, whereas both, CAR and ARG supplements appeared to promote muscle maturation, as indicated by a greater lactate dehydrogenase-to-citrate synthase ratio. Whether this has a long term effect resulting in improved growth rate in the starter, grower and finisher period needs further studies.
In the second trial the same diets were used as in the previous trial. However, as feed intake in the previous trial was on average about 20% lower than expected, piglets were offered permanent ad libitum access to the CON, ARG and CAR milk replacer via a milk cup system, with the intention of increasing feed intake to levels close to those expected in suckling piglets. There were a few responses in indicators of protein synthesis of the light weight piglets from large litters to L-carnitine (enhanced mTOR phosphorylation in the light portion of the STM) and L-arginine (expression of a gene involved in protein synthesis in light portion of the STM). However, different from previous studies, the concentrations of L-arginine and L-carnitine chosen to be supplemented to a quite common type of milk replacer did not elicit the expected effects on growth performance of early-weaned low weight piglets even when offered at ad libitum access. One explanation for the lack of effect could be that the requirements are not up to date with current genotypes of low weight piglets from hyperprolific sows. Another, and more likely, explanation is that the milk replacer was not only lacking arginine, but possibly also other amino acids like lysine in comparison with sow milk. Thus, the approach of only supplementing arginine would lead to an amino acid imbalance. Further research should focus on a more holistic strategy concerning the entire amino acid composition of milk replacers. The high survival rate observed in this study suggests that applying rescue decks can be an important measure to reduce losses of piglets from large litters during the mid to late nursery phase.
The third study was divided in two phases, one phase investigating the growth performance and muscle development of artificially reared low birth weight piglet from 7-28 days of age fed either an unsupplemented milk replacer (CON) or a milk replacer supplemented with a combination of L-arginine and L-carnitine (CarArg). In the second phase, selected CON and CarArg piglets of the first phase were used to investigate the possible effect of artificial rearing and CarArg supplementation on post-weaning growth performance, carcass composition and meat quality traits at 170 days of age. The combination of the dietary supplements L-carnitine or L-arginine and artificial rearing did not promote pre- or post-weaning growth of low birth weight piglets from large litters. However, compared with conventionally reared low birth weight pigs from litters with markedly reduced litter size from day 7 to 28 of age, the overall growth performance and final slaughter weight did not differ. Also carcass composition and meat quality traits were not affected by dietary treatments or rearing strategy during nursery. Hence, excess number of low birth weight piglets from large litters can be reared artificially without compromising the overall production efficiency. However, as supported by the variation of slaughter weight among all pigs across dietary treatments and rearing strategy, it seems that birth weight alone is a poor predictor for the pre- and post-natal growth potential. For this reason future research should focus on selecting piglets for artificial rearing in the nursing period, based on easy recognizable morphological traits that show strong correlation to post-natal growth.

Work Package 3

Understanding gut-related factors associated with feed efficiency in chickens
Using similar experimental protocols, the environmental effect on feed efficiency-related differences in the intestinal microbiota and function was studied in chickens reared at two research sites (in Austria (Vetmeduni) and Northern Ireland (Agri-Food and Biosciences Institute). The best and worst feed efficient chickens were selected. Feed efficiency values using the residual feed intake as a metric for feed efficiency were similar between locations within the same feed efficiency group, and differed by 450 g between ‘good’ and ‘poor’ feed efficiency in females and male chickens. Despite a similar core microbiome, the bacterial composition associated with low and high feed efficiency was different between the two locations and between females and males. Only few bacterial phylotypes were consistently associated with feed efficiency across locations. Two Lactobacillus crispatus-phylotypes in the feces of females and a Ruminococcus phylotype in the ceca of males were negatively correlated with feed efficiency. By contrast, spore-forming Dorea and starch-degrading Turicibacter phylotypes were positively correlated with feed efficiency in males and thus may serve as potential marker phylotypes for chicken´s feed efficiency. The rearing environment also greatly affected intestinal size and function, whereas feed efficiency related differences in intestinal size, structure and function were mostly location-specific. Feed efficiency was more correlated with intestinal parameters when the feed conversion ratio was used as feed efficiency metric than with the residual feed intake. Serum metabolites, such as serum uric acid and cholesterol, were identified to be possible markers for feed efficiency in chickens.

Understanding gut-related factors associated with feed efficiency in pigs
Due to a disease outbreak at Hermitage Genetics early in the project’s life and although not initially planned, two independent studies emerged from the pig work relating to the objective to understand gut-related factors linked to feed use efficiency. The first study involved eighty one pigs ranked at a research centre in Ireland (Teagasc) on RFI between weaning and day 126 post-weaning, where 32 were selected as the extremes in RFI (12 low, 10 medium, 10 high). Intestinal microbiota diversity, composition and predicted functionality were assessed by 16S rRNA gene sequencing. Although no differences in microbial diversity were found, some RFI-associated compositional differences were revealed, principally among members of Firmicutes, and predominantly in faeces at slaughter (albeit mainly for low abundance taxa). In particular, microbes associated with a leaner and healthier host (e.g. Christensenellaceae, Oscillibacter, Cellulosilyticum) were enriched in low RFI (more feed efficient) pigs. Differences were also observed in the ileum of low RFI pigs; most notably Nocardiaceae (Rhodococcus) were less abundant. Predictive functional analysis suggested improved metabolic capabilities in these animals, especially within the ileal microbiota, which was corroborated by higher ileal isobutyric acid concentrations. Taken together, these data indicate differences, albeit relatively subtle, within the intestinal microbiota of low RFI pigs compared to their high RFI counterparts, demonstrating a possible link between the intestinal microbiota and FE in pigs.

The second study investigated whether FE-associated enterotypes and selected FE-associated physiological traits in pigs were consistent across Teagasc ([two batches of pigs; differing health status, ROI1 (low health) and ROI2 (high health)], AFBI, and Vetmeduni, where differences in genetic, dietary and management factors were minimized. Pigs (n=369) were ranked on divergence in residual feed intake (RFI, a metric for FE) and 100 extremes were selected in total (50 high RFI and 50 low RFI) across geographical locations for analysis of the intestinal microbiota and a range of FE-associated traits. Intestinal microbial diversity, determined by 16S rRNA amplicon sequencing, varied with geographical location, health status and intestinal site, but not by RFI. For the high health batch of pigs (ROI2), low RFI (more feed efficient) animals had greater species diversity in the ileal and caecal digesta. In addition, Lentisphaerae, Mucispirillum, Methanobrevibacter, Ruminoccocaceae, RF16 and two uncultured bacterial taxa were more abundant within the faecal/caecal microbiota of low RFI pigs in two geographic locations and / or the two batches from the Teagasc. In particular, Lentisphaerae was correlated with FE. These taxa are major contributors to carbohydrate metabolism, which was reflected in the functional predictions. Low RFI pigs had lower faecal concentrations of total, valeric, butyric, and propionic volatile fatty acids, which correlated with the increased abundance of members of Lentisphaerae. None of the other physiological traits measured varied according to RFI, except salivary cortisol which tended to be lower in low RFI pigs. It was concluded that the rearing environment, health status and intestinal site greatly impacted the pig gut microbiome, which, in turn, presents challenges when identifying consistent reliable microbial biomarkers for FE in pigs. However, several FE-associated enterotypes were common across batches and geographic location and related to a potentially “healthier” and metabolically more capable microbiota. These taxa could therefore be employed as biomarkers for FE and may merit consideration for use as probiotics or targeting by dietary means as a strategy for improving FE in pigs in the future.

Vetmeduni analysed samples from the cross centre study referred to above. Here the objective was to investigate the differences in visceral organ size, intestinal morphology, mucosal enzyme activity, intestinal integrity and related gene expression in low and high RFI pigs when reared at the three different geographical locations (Austria; Northern Ireland; Republic of Ireland) using similar protocols. It was concluded that geographic location had a stronger greater effect on visceral organ sizes and intestinal structure and functionality than the RFI rank of the pig.

In Teagasc, the growth data from pigs collected as part of the first trial described above for this research site was used to quantify the inter-relationships between a range of different short-term (feed) efficiency measures in growing pigs and characterize pigs divergent for a selection of these measures. Growth-related metrics available included mid-test metabolic weight (BW0.75) energy intake (EI), ADG, as well as both live-animal backfat (BF) and muscle depth (MD). Ratio efficiency traits derived included energy conversion ratio (ECR), Kleiber ratio (KR) and relative growth rate (RGR). Residual energy intake (REI) and residual daily gain (RDG) were also quantified; a dual index of both REI and RDG was calculated and termed residual intake and BW gain (RIG). A new residual trait, termed residual mid-test metabolic weight (RMW), was also calculated which represented the residuals from least squares regression of BW0.75 on both EI plus ADG. Considerable inter-animal variability existed in all metrics evaluated. Male pigs were superior to females for all metrics evaluated (P < 0.001) except for both BW0.75 and EI where no gender differences were evident. Feed efficiency metrics of individual animals improved as birth BW increased (P < 0.05) except for RGR where the contrary was observed. Phenotypic correlations between most growth and feed efficiency metrics were strong to moderate (P < 0.05 when compared against no correlation) although most differed from unity. Results suggest that each feed efficiency trait measures a different aspect of efficiency although improvement in one of the traits would, on average, lead to improvements in the others.

Development of synbiotics for in-ovo use in broiler chickens
At IBB (Poland), to develop synbiotic formulations for efficient feed conversion in chicken, several Lactobacillus strains from the IBB collection were tested for their carbohydrate catabolic potential, enzymatic profile, growth at various conditions (temperature, sodium chloride, acid and bile salts concentration), adhesion under static conditions and antibiotic susceptibility. Finally, two bioactive synbiotic formulations were prepared, containing Lactobacillus. plantarum IBB3036 + prebiotic lupin Raffinose Family Oligosaccharides (SYN1) and SYN2 (Lactobacillus salivarius IBB3154 + prebiotic belonging to transgalactooligosaccharides, Bi2tos, Clasado Ltd.).
To evaluate doses of synbiotics for the in ovo experiments, bench studies were completed at UTP (Poland). Subsequently, PCR was used to detect bacteria in chicken droppings after in ovo injection. Based on results of PCR assay and analyses of embryo mortality (conducted by UTP), for further experiments, dosage containing the highest concentration of both components of bioactives were chosen. These formulations were injected in ovo to determine their effect on gut microbiota of post-hatch broiler chicken. Experiments on chickens and quantitative PCR showed that in ovo administration of symbiotic formulations results in a small reduction of Enterococcus faecalis detected in droppings in contrast to the control group. Moreover, both strains were detected in 3-day-chicken fecal samples, but only L. salivarius IBB3154 was able to persist and tended to increase in chicken’s gastrointestinal tract up to 6 weeks after hatch.

Research results obtained at the UTP site (Poland) showed that in order to ensure a persistent effect of a synbiotic formulation on the chicken’s gastrointestinal tract, colonisation of beneficial bacteria and improved production, the bioactive must be administrated under precisely controlled conditions on the 12th day of egg incubation. Two novel synbiotic formulations were designed and optimised in vitro (IBB) as described above (S1 – L. salivarius with galactooligosaccarides (GOS) and S2 – L. plantarum with raffinose family oligosaccharides (RFO)). The recommended dose for injection was 105 cfu/egg of probiotic and 2mg/egg of prebiotic. Post hatching, 2,040 male chickens were reared (680 individuals/group split in 8 replicate pens). Body weight and feed intake of the chickens were not changed by the treatment, but both synbiotics had beneficial effects on the overall status of the organisms, as indicated by low mortality and improved production parameters. Microbial populations of Lactobacillus spp./Enterococcus spp. in the ileum were higher in synbiotic groups vs. control. S1 caused significant up-regulation of IL6, IL18, IL1β, IFNγ and IFNβ in the spleen on day 21. In cecal tonsils, S1 caused significant down-regulation of IL12, IL8 and IL1β on day 42 and IFNβ on day 14 (P < 0.05). Out of the two synbiotics tested in this work, S1 seems to be more potent in establishing down-regulatory pattern of the immune-related gene expression in the GALT and beneficial shifts in the gut microbiota composition. As for physiological traits, S2 induced stronger physiological changes than S1. It increased trypsin activity in the pancreas and elevated lipase activity in duodenal content. Synbiotics strongly decreased intestinal expression of mRNA for incretins, which coincides with altered activity of digestive enzymes. Synbiotic S2 had a marked effect on total lipid content (lower) and FA composition with a positive effect on nutritional properties of chicken meat.
In general, careful pre-selection of the synbiotic and its validation is crucial to obtain desired effects on the animal upon in ovo treatment. Since in vivo examination of the bioactive properties of synbiotics is time-consuming, labor-intensive and involves large numbers of animals, we recommend that in vitro screening of synbiotics prior to in vivo studies should be conducted first to evaluate potential biological properties of such synbiotics and their impact on the host.

Post-hatch treatments for broiler chickens to promote feed use efficiency
At Vetmeduni (Austria), to promote chicken’s feed efficiency post-hatch via a modulation of the gut microbiota chickens received an inoculum that was prepared from faeces of feed-efficient chickens on day 1, 6 and 9 of life. The inoculum mainly comprised Enterobacteriaceae and Turicibacter, which were both identified as important bacterial groups for feed efficiency in chickens in the previous chicken experiment. However, effects of the inoculum on the intestinal bacterial profiles and chicken’s feed efficiency were small. In females, the inoculum promoted feed intake and growth but did not change the chicken’s feed efficiency. Although effects on intestinal size, structure and function were small in both sexes, the inoculum seemed to change the feed efficiency-related differences in total gut weight in females, structure (e.g. caecum length in females) and intestinal permeability in the jejunum of males but without changing the chicken’s feed efficiency. The other intervention strategy that was tested to ameliorate chicken’s feed efficiency via the modulation of the gut microbiota was that chickens were fed ad libitum or restrictively (~95% of ad libitum feeding). From the results it can be concluded that restrictive feeding from day 9 of life may be one option to improve the feed efficiency of low feed efficient chickens, particularly in females. Restrictive feeding had a marked effect on the abundance of bacterial families in ileal and caecal digesta and faeces of females and males which may have contributed to the improved feed efficiency. In this experiment, an increased abundance of Lactobacillaceae in ileal digesta of restrictively fed chickens seemed to be advantageous for the feed efficiency of the host; despite the fact that Lactobacillaceae were previously associated with low feed efficiency. Although intestinal size, structure and functionality were mostly similar, observed feed efficiency-related differences in gut structure (e.g. ileal crypt depth in males) and size (e.g. liver weight in females) were different when chickens were fed restrictively compared to those fed ad libitum, which may have contributed to improve the feed efficiency of low feed-efficient chickens.

Treatments to promote feed use efficiency in pigs
Our results from earlier research in this work package showed that the porcine intestinal microbiota is linked with feed efficiency (FE) in pigs. Improving FE by manipulating the intestinal microbiota may therefore be possible. An experiment was conducted at Teagasc with the objective to investigate changes in intestinal microbiota and growth in pigs following faecal microbiota transplant (FMT) in sows and/or their offspring with faecal extracts from highly feed efficient pigs. Twenty two sows were assigned a) Control or b) FMT during gestation. Offspring from both sow treatments were assigned, within litter, to i) Control, ii) FMT at birth, or iii) FMT at four times. At weaning, 84 piglets (n=28/offspring treatment; n=42 /sow treatment) were individually housed up to slaughter (~155 days old), and 36 pigs were selected (6/treatment) for sampling. Offspring body weight at slaughter age was reduced due to FMT, potentially due to changes in the intestinal microbiome. Bacterial diversity was higher, and many compositional changes were observed, mainly in the ileum. Offspring had a decreased abundance of beneficial microbes such as Faecalibacterium, Oribacterium, Butyricimonas and Terrisporobacter. In this study, FMT conducted to manipulate the intestinal microbiota in order to improve FE, influenced intestinal microbiota composition in a way that negatively impacted lifetime pig growth and associated physiological parameters.

In a second experiment conducted at Teagasc, faecal microbiota transplantation (FMT) was performed in sows at days 70 and 100 of gestation [control sows (CON, n=11), FMT sows (FMT, n=11)]. Subsequent offspring were assigned to one of two treatments i) Control pigs (con, n=67) and ii) Inulin added to the diet of weaned pigs (at 28 days of age) for six weeks (2% for 3 weeks and 3% for 3 weeks; inu, n=65). The combined strategy reduced the weight of offspring at slaughter, although it was associated with a low RFI value (a metric for feed efficiency). Shifts in the intestinal microbial diversity, composition and predicted functionality were observed throughout the lifetime of offspring, due to FMT in sows and the addition of inulin to the offspring diet. At 130 days old, all groups had significantly higher diversity (Shannon index) than the CON/con. Mainly at weaning and at the age of 100 and 130 days old, many changes in important bacterial populations (e.g. Bifidobacterium, Butyricimonas, Streptococcus, Chlamydia and Faecalibacterium) occurred. However opposite effects at sow treatment compared to offspring treatment were observed. Ileal acetic acid concentrations increased in FMT/inu compared to FMT/con counterparts while propionic concentrations decreased in CON/inu offspring compared to the CON/con group. Notably, offspring from FMT sows (regardless of their own treatment) had lower caecal butyric acid concentrations compared to offspring from CON sows. These pigs also had more goblet cells in the jejunum, and inulin fed offspring had a higher ileal villus height:crypt depth ratio, compared to their control counterparts. Furthermore, offspring blood haematological parameters including white blood cells and granulocytes and blood urea nitrogen concentrations were reduced due to the addition of inulin to the diet, whereas the mean platelet volume was increased. Duodenal gene expression showed that the addition of inulin to the diet up-regulated genes linked to glucose (GIP, GLP1) and VFA (SMCT) homeostasis in comparison to the control group suggesting healthier metabolic status.

Work package 4

In order to identify the genetic architecture of feed efficiency (FE) in pig and chicken, holistic genome-wide association studies (GWAS) were conducted in both monogastric species. Therefore, 862 chicken of a Cobb line (partner 17) and 1296 boars provided by Hermitage (partner 5) were genotyped at about 60000 DNA markers (single nucleotide polymorphisms (SNPs) using SNP-chips; 60K Illumina beadchips) by partner FBN (Germany). In chicken, phenotypic data covering body weight at 36, 39 and 46 days (BW36, BW39, BW46), body weight gain between 39-46 days (BWG), feed intake between 39-46 days (FI), feed conversion ratio (FCR) were analysed; in pigs the investigated traits comprised feed conversion ratio (FCR), daily occupation time (DOT), daily feed intake (DFI), daily number of feeder visits (DFV) and daily consumption rate (DCR) as well as other production traits like growth rate (D110) and percent lean (PCL). The heritability for feed conversion was low in chicken (h²=0.1) and moderate in pigs (h²=0.3). The difference among species is due to the fact that for chicken the analyses cover the upper 20% of individuals in terms of growth performance only. Accordingly estimated marker effects were generally higher in pigs ranging from 3 to 7% of the phenotypic variance compared to broilers where marker effect were almost all close to 1%.

Genome-wide association approaches were employed by FBN which analyse marker by marker (single-marker regression analysis), or considering windows covering several markers at a time (multi-marker Baysian analysis). The analysis of genomic regions important in defining FE within important chicken and pig breeding lines revealed in total 27 and 100 QTL regions for all analyzed traits in the commercial Cobb broiler line A and the Maxgro terminal sire line population, respectively. Regarding the genetics of FE-traits in chickens, single-marker and multi-marker GWAS revealed quantitative trait loci (QTL) on 13 different chromosomes. The integration of both approaches resulted in 7 overlapping QTL regions that showed contribution to the genetic variance of the trait > 0.5%, and contained at least one significantly associated SNP (–log10[p-value]≥4.3). Also for pigs, the analyses of traits related to FE and feeding behaviour revealed several QTL regions on 10 chromosomes, with partial overlaps for some traits. The integration of single-marker and multi-marker analyses for carcass traits revealed 11 and 10 regions showing significant association with D110 and PCL, respectively.

In addition to the GWAS entropy analyses were also applied by Poznan University (Poland). This is a recently used method in animal genetics in order to detect SNP being most informative with regard to the production traits and to estimate interactions between the most informative pairs of SNPs. The analyses revealed a broad distribution of informative SNPs over most chromosomes in chicken and pigs. For example for BW36 in chicken, 154 SNPs were informative and enabled to name positional candidate genes. For FCR in pigs, the majority of ‘most informative SNPs’ exceeding the 5% threshold were located on chromosomes: 1 (23 SNPs), 2 (14 SNPs) and Z (26 SNPs). In these regions 52 genes were found.

Towards the development of tools for breeding and to develop a cost-effective solution for the implementation of GS in pigs, a set of SNP-markers was selected which enables the establishment of a low density genotyping array. The set combines the priority list of SNP markers, i.e. 88 significant markers with the highest contribution to the genetic variance of FCR, minor allele frequency > 5%, high call rates and in linkage equilibrium, and a set of equally distributed informative markers covering all chromosomes. The final set of markers was further tested in a cross-validation approach. Moreover, the accuracy of breeding value prediction was compared for sets with different SNP densities (3K, 6K, 12K, 22K, 60K) in order to assess the implementation of the results towards GS for improved FE in pigs. The analyses revealed that a higher number of included SNPs results in a larger accuracy of estimated breeding values and consequently larger genetic gain. Due to the fact that the priority list comprises markers with highest contribution to the genetic variance, a 96 SNP chip showed correlations in the cross-validation of approximately 0.45. The 3k and 6k SNP sets performed substantially well compared to the low density sets. Consequently, a combined approach was recommended to the partner (Hermitage) for a cost-effective low density chip covering all chromosomes with 3,000 SNPs including the 88 priority SNPs derived from QTL regions.

From a biological point of view, FE largely depends on the digestion and absorption of nutrients in the gut, partitioning and primary metabolism in liver, and superior mechanisms to orchestrate resource allocation for maintenance, growth, physical activity, and thermoregulation. Moreover, there is an increasing body of evidence that the genetic selection towards FE in monogastric species employs an array of molecular mechanisms which mainly rely on the routes to utilize macronutrients. In this context, different animal trials including closely related finishing pigs and broiler chickens which considerably differ in their feed efficiency (FE) were investigated regarding their transcriptomic profiles of relevant tissues. Local and systemic effects on nutrient metabolism were deduced from transcriptome and fatty acid profiles of liver tissue as well as blood parameters and metabolites. Based on differentially abundant transcripts between low and high FE groups in liver, canonical pathways like integrin signalling and lipid and carbohydrate metabolism were shown to be affected by divergent FE. Specifically, the molecular alterations of lipid metabolism emerge a pattern of an overall reduced hepatic usage of fatty acids in high FE animals. Complementary analyses at the systemic level exclusively pointed to an elevated level of triglycerides in high FE pigs. Fatty acids of liver tissue showed considerable lower concentrations of saturated and poly unsaturated fatty acids in the high FE group. Thus, FE traits were reflected by hepatic profiles of fatty acids and gene expression in pigs. These results suggested that strategies to achieve FE favour a metabolic shift from energy storage towards energy utilization and mobilization. Moreover, improvements in FE traits might be driven by an intensified cellular infrastructure as suggested by transcriptome data. The genes PDK2 and FYN were identified as interesting candidate molecules that are suggested to play a central role in mediating FE in hepatic tissue. Accordingly, these genes are putative biomarkers indicating the contemporary utilisation and partitioning of nutrients.

In fact, the protein encoded by PDK2 is known to inhibit the hepatic pyruvate dehydrogenase complex which acts as a gatekeeper in carbohydrate metabolism mediating a key metabolic switch from glucose utilization towards increased fat metabolism. The second proposed candidate molecule FYN was identified as one of the transcripts exhibiting the highest responsiveness to FE traits in liver and was also indicated by GWAS performed within ECO-FCE. Based on the functional annotation of FYN, it is involved in cell growth, cell adhesion, integrin-mediated signalling, cytoskeletal remodelling, cell motility, and immune response. Moreover, FYN is supposed to be a major factor regulating fatty acid utilization and glucose homeostasis via effects on AMP-activated protein kinase activity. Thus, it is conceivable that FE traits represent a metabolic shift from energy storage towards energy utilization and mobilization. Indeed, characteristics of lipid metabolism such as lipid and fatty acid profiles, enzyme activities, and associated transcript abundances have been reported to be altered in FE divergent animals.
Therefore, transcriptomic analyses within ECO-FCE further focused on molecular alterations in muscle and adipose tissue (subcutaneous) of FE-divergent pigs. In the adipose tissue of pigs, holistic analyses conducted by partner FBN (Germany) revealed that the differentially expressed genes were enriched for functions such as cellular movement, growth & proliferation, haematological system development, immune cell trafficking and cell death & survival. Together with the observed repression of endothelial growth functions this suggests an impact on adipose tissue development and extracellular matrix formation in more efficient pigs. However, the most significantly enriched pathways in adipose tissue were complement system & acute phase response signalling. In accordance with findings in liver tissue as well as observations of an altered muscle-fat balance, transcriptional alterations in muscle tissues were classified into three categories: accumulation of fatty acids, adhesion of connective tissue, and apoptosis. Nevertheless, the adaptive responses at the level of the muscle also comprise efforts to maintain tissue integrity suggesting altered immune functions as a central component in different relevant tissues towards improved FE. High FE pigs may display a more rapid and efficient inflammatory response. Overall, findings are in accordance with phenotypic observations of reduced development and formation of adipose tissue in more efficient pigs and depict putative links between strategies to cope with inflammation and FE.

Moreover, an integrative approach which combines weighted gene co-expression network analyses and multivariable methods integrating transcriptomic and phenotypic information was used for liver and duodenum samples of trials conducted at IRTA (Spain). These analyses revealed interesting tissue-specific candidate genes (e.g. GADD45G, MAP3K8, NFKBI or SGK1 in liver; IL22, SIN3A, NQO1, CSMD2, CDC37L1, PSAT1 or LECT2 in duodenum), as well as 12 genes that were common to both tissues (e.g. HSPH1, BAG3, DNAJA1, STIP1). Some of these identified candidate genes had been previously detected as associated to feeding behaviour or to feed efficiency in other species. Pathway analyses indicated that these genes take part in a wide variety of physiological and biological events. Again, this also comprise immune-related functions such as inflammation, immune response and heat shock protein binding. The functional information of the holistic expression analyses has been successfully integrated with genotype information towards the development of an improved genomic selection approach, the genomic feature best linear unbiased prediction (GFBLUP) model developed by partner Aarhus (Denmark).

RNA-sequencing analysis of broiler chicken samples performed by FBN revealed 172, 120 and 81 differentially expressed genes (DEG) in duodenum, jejunum and ileum between FE-divergent broilers. For liver, leg, and breast muscle, 159, 186, and 54 DEG were identified. Thus, the transcriptomic analyses highlighted the specialisation of most FE-related tissues/organs but, nevertheless, showed that they are also multi-functional affecting an array of distinct biological processes. Specifically, enriched molecular pathways for gut samples were related to growth, immune function, glucose uptake, transport, and lipid metabolism. In fact, specific transporters such as members of the solute carrier transporter family were affected by FE traits. Moreover, genes found to be differentially expressed between experimental groups are involved in pathways relevant to nutrient utilisation and are potential candidate genes for improved FE.
Therefore, the molecular pathways and bio-functions revealed by transcriptomic analyses in chickens showed similar pattern compared to the analyses in pigs. In liver, muscle, and adipose tissue this primarily comprise effects on the muscle-fat balance which might have their origin in a shift in the utilization of macronutrients. This consistent pattern is also in accordance with long-term FE-selection experiments performed in pigs. In addition, the analyses of almost all investigated tissues provide several lines of evidence for an important role of immune functions for divergent FE. This might imply an adaptation of high FE animals in terms of resource allocation under controlled housing conditions. Therefore, the resilience of high FE under different (e.g. challenging) conditions is of high scientific importance. The results of transcriptomic analysis in chickens further provide evidence that sex-specific effects dominated FE-specific effects. Consequently, to estimate the impact of FE traits on gene expression and to understand related physiological gastrointestinal features, the sex-dimorphism in FE-divergent chickens need to be taken into consideration.

Beside the efficient conversation of macronutrients by monogastric species, efficiencies for micronutrient such as phosphorus (P) also play a considerable role towards the production of high-quality animal-derived food considering environmental aspects. In chickens, two different trials with 96 male broiler chickens per trial were performed at IRTA to investigate if chickens under an early restriction in dietary P or methionine have the ability to better utilize these specific nutrients later in life. For the P conditioning trial, an early conditioning and a late P restriction significantly increased the expression of the major sodium-P cotransporter (SLC34A2) in different parts of the small intestine as revealed by targeted expression analysis. In a holistic approach, further putative candidate genes mediating short and long-term adaptations to the conditioning were identified, such as DIP2C, EAF2, SPON2 and MFSD2A in P-conditioning, and CHST4, HSD17B12, KCNK16, MET, DOCK3, NOV in methionine-conditioning. Differential expression was much higher when comparing animals under nutritional conditioning with control animals: 46 and 48 DE genes in P and methionine conditioning experiments. This suggests slight but relevant adaptations at the transcriptional level following a nutritional conditioning.

In pigs, molecular adaptations to divergent P levels were monitored in peripheral blood mononuclear cells (PBMC) by partner FBN. PBMC represent physiological and environmental-sensitive transcriptional shifts in other tissues and organs. Diet-specific serum hormone and mineral alterations display responses regarding variations in P supply, that indicate a regulatory involvement of major factors contributing to the maintenance of P homeostasis. The endocrine system revealed diet-specific organismal efforts to minimize P losses or to induce a negative P balance, respectively. Therefore, the dietary challenge revealed considerable developmental plasticity in pigs. Additionally, the transcriptional shifts of pathways related to acquired and innate immunity imply a trade-off between micronutrient supply and immune system. Genes found to be differentially expressed due to variable P-supply are involved in pathways relevant to P-utilization and are potential candidate genes and biomarkers for improved P-efficiency. Furthermore, analyses in jejunum, colon, and kidney revealed various genes being involved in adaptive responses to maintain the mineral homeostasis. Transcripts associated with the complement system and NFAT signaling underline the involvement of P in the immune system. The results suggest that a low P diet might be part of a scenario towards improves resource efficiency and reduces environmental burden. However, the potential impact of P supply on immune features should be considered in animal husbandry and policy/governance aspects.

Pork and chicken quality encompasses a diverse array of characteristics relating to eating quality, nutritional level and technological performance, while broader definitions can also include shelf life and safety. Quality can be influenced by production factors such as breed, feeding strategy, rearing conditions and slaughter age/weight, and over the past 40 years genetic selection for increased lean growth rate and feed efficiency has been associated with reduced intra-muscular fat content (IMF), altered post-mortem pH profile, resulting in antagonistic effects on tenderness, juiciness, flavour and processability and/or the ability of the muscle to retain water in pork and to a somewhat lesser extent in poultry. In ECO-FCE, it was considered important that in positively driving selection for feed efficiency and optimisation of management strategies for improved efficiency, implications for product quality should be assessed. For this reason, meat quality was analysed by Teagasc, IRTA and UTP in numerous trials reflecting genetic and management strategies, and in pork and chicken, conducted at three diverse sites i.e. Teagasc (pigs), IRTA (pigs and chicken) and UTP (chicken). In all cases, a broad range of meat quality parameters were investigated, including trained panel assessment of eating quality traits. Analyses were standardised as far as possible across centres so that results are comparable. Results were similar across the research partners. The main outputs indicate that certain aspects of meat quality were affected significantly but to a minor degree by intrinsic divergence in FE and by selected management interventions. Importantly, none of the resultant meat was of unsatisfactory quality.
High FE pigs showed differences in post-mortem evolution of pH and reduced water holding capacity in one study. These unfavourable associations could potentially be explained by greater glycogen content in the muscle of high FE pigs. However, ultimate pH was still in the acceptable range and none of the samples would be described as impaired in quality. For example, no muscle was identified as exhibiting specific defects such as being pale, soft and exudative (PSE). The implications of reduced water-holding capacity could be significant for processors, in terms of yield, and this trend should be considered when driving efficiencies further than in the present project.
In terms of carcass quality, better feed converters had leaner carcasses with greater muscle content. However, as evident in past studies, and in selection programmes for lean growth, this was also associated with significantly less intramuscular fat content. For example, while both low and high FE groups raised in Teagasc (Ireland) were generally lean, high FE pigs had less intramuscular fat content compared to low FE pigs. This effect was also observed at IRTA (Spain) and is in accordance with results obtained from transcriptional analyses of liver, muscle and adipose tissues. While this could potentially impact on flavour of resultant meat, our trained sensory panel data indicated no effect on sensory parameters closely related to intra-muscular fat content, suggesting the observed difference in fat content might not be of sufficient magnitude to be discernible by consumers. On the positive side, the health profile of the fats present was more favourable for the high FE pigs, compared with low feed efficient animals. With regard to texture, results for shear force and sensory texture varied between studies. Significant effects on sensory tenderness were observed in Teagasc, but not in IRTA. Longissimus samples from high FE samples achieved slightly lower scores for sensory texture parameters compared with low FE.

In poultry, effects on meat quality were small, however improving feed efficiency through diet modification increased intramuscular fat content while it also altered meat colour and pH. Furthermore, in ovo synbiotic studies demonstrated no major impact on meat quality traits.
To conclude, improvement in feed efficiency revealed meat quality characteristics, in particular the key traits of tenderness, juiciness, flavour and water-holding capacity generally within the normal range for the production of acceptable quality pork and poultry. It would be beneficial to carry out consumer panel studies to assess the acceptability of meat from a hedonic point of view.

Work package 5

The foreground of this section composed of a set of tools that aim to assist the farmer/breeder in defining the best genetic strategies to improve efficiency and reduce environmental pollution.
These tools are described in detail below:
Genetic modelling tool
In order to conduct research on the genetic and biological basis of the feed efficiency complex, phenotypic and genetic datasets on pigs and chickens need to be established. The phenotypic data of longitudinal feed intake during the growing period of pigs, along with single recorded growth traits, were made available. The genomic data allowed construction of a reference population for implementing genomic selection. The statistical method of longitudinal random regression was developed and tested on mink as a model organism, in the first study which revealed that the genetic dynamics of feed efficiency can be dissected utilising the random regression method. The random regression method was also used to model the body weight growth of chicken over generations. This method revealed that body weight growth in three age points in chicken have different genetic backgrounds, with later ages having larger genetic variance compared to early stages, and showed that genetic variance did not decline even after several generations of selection.

Raw longitudinal feed intake visit data in pigs were quality controlled for errors using specially designed algorithms. The procedure involved obtaining weekly averages of feed intake and feeding behaviour traits including feeder occupation time, number of visits to the feeder and feeding rate. Residual feed intake, as a partial measure of feed efficiency which explains the proportion of feed intake that is independent from production traits, was used to explain the genetic background of feed efficiency. Random regression analysis of residual feed intake and feeding behaviour traits revealed the varying genetic background of feed efficiency. Modelling the dynamics of feed efficiency can enhance the genetic improvement of feed efficiency by considering the change in feed intake capacity of pigs. The feeding behaviour trait of occupation time showed that spending more time at the feeder results in lower feed efficiency due to higher feed consumption in pigs. Furthermore, lower numbers of visits to the feeder and lower rates of feed consumption per visit can improve the efficiency of nutrient digestion in pigs. Feeding behaviour traits can be included in routine genetic evaluation to further improve the genetic progress of feed efficiency by better accuracy of prediction for feed efficiency.

Traditionally feed efficiency is defined by the feed conversion ratio (feed intake over the body weight gain), which has complex biological and statistical properties. However, to circumvent the problems of ratio traits, residual feed intake has been proposed as a better measure to determine an animal’s feed efficiency. In ECO-FCE, a Bayesian statistical methodology to propose guidelines for measure of feed efficiency in breeding programs and scientific research was developed. The results showed that the ratio trait of feed conversion ratio has unfavourable statistical properties, such as unbalanced selection pressure on growth traits involved in breeding programs; while, residual feed intake measured using genetic (co)variance matrix allows for improvement on that proportion of feed intake that is not used for growth and body composition. This methodology is applicable to any ratio and linear traits across different species.

In routine breeding programmes feed intake is recorded longitudinally across test periods while growth traits of body weight gain and lean meat percentage are recorded as single traits. To utilise longitudinal feed intake records combined with single recorded growth traits to obtain longitudinal genetic residual feed intake as a measure of feed efficiency requires novel statistical methodology. In ECO-FCE, we developed a novel Horizontal method that can be used to analyse repeated records together with single recorded traits through specific modelling of the environmental (co)variances. The Horizontal model was implemented on three diverse Danish pig breeds. The Horizontal model allows modelling the change in feed intake capacity during growth. Clear breed differences between Duroc and dam line breeds in the genetic parameters of the feed efficiency complex were illustrated throughout the period of growth. Feed intake and genetic RFI had varying genetic backgrounds during the growing-finishing period in pigs. Defining RFI as genetically independent of production traits led to easily interpretable breeding values.

Large scale genomic information is increasingly becoming available in livestock. The methodologies are required to facilitate integration of molecular information, such as important genomic regions associated with traits of interest, with whole genome information. In ECO-FCE, we developed novel genomic feature models that generate or utilise knowledge of the biological mechanisms underlying trait variation and integrate them with whole genome information. The genomic feature method showed increased predictive ability of genetic merits of animals when significant genomic regions were included compared to standard methods. This approach was tested over genetic residual feed intake as a measure of feed efficiency along with production traits and showed that the genomic feature model increases the accuracy in predicting the genetic merits of elite animals in the population specifically for complex traits such as feed efficiency.

3-in-1 ecological calculators
Animal performance as well as environmental inputs are the main drivers of environmental outputs of nitrogen (N), phosphorus (P) and green house gases (GHG’s). An ecological calculator which predicts the main environmental outputs (N, P and GHGs) from monogastric production systems was designed to compare and contrast the main effects of different experimental treatments on the excretion of N and P and emissions of GHG’s. The final 3 in 1 calculators (one for pigs and one for broiler chickens) calculate N and P excretion based on output figures of the animals performance (feed intake and growth rate) as well as dietary composition. They calculate N excretion by assuming that N retention was (ADG × 0.16) / 6.25 for pigs (Whittemore et al., 1988) and (ADG × 0.175) / 6.25 for broiler chickens (Gheorghe et al., 2013). N excretion is then calculated by subtracting N retention from N intake. For determination of P excretion, the equation for P retention was ‘Bodyweight × 0.95 × 0.0058’ for pigs, and ‘Bodyweight × 0.0058’ for broiler chickens, which assumes a figure of 0.58% P retention of the slaughter weight was used. P retention was subtracted from P intake to estimate P excretion. Both N and P intakes were based on individual dietary stages when dietary information was provided.
The output from the GHG calculators represented the sum of the emissions of methane (CH4) from the animal and manure, and emissions expressed in terms of the CO2e from the feed including emissions from feed processing. The main input values were animal finish weight, average daily feed intake and the total days to slaughter.
The final pig and broiler chicken calculators were both excel based and presented in separate files. The calculator interface allowed users to input the weight of the animals at different stages of growth with feed intake details as well as feed composition details. Default values for birth weight, average days to slaughter, average feed intakes and bodyweights at starter and grower stages, days spent on starter and grower diets and dietary composition of CP and P were included and could be used when such data were not available. A default value for the GHG emission factor of feed was also included, but advice was provided on how more accurate emission factors for the specific diets used in the trials could be attained (i.e. through FeedPrint ( This advice was provided since emissions factors could vary greatly depending on the raw ingredients used, e.g. a simple soya bean meal based diet had an emission factor of 0.453 kg CO2e per kg feed, whereas a diet where soya was replaced with wheat DDGS and RSM had an emission factor of 0.398 kg CO2e per kg feed. The input of more refined data in terms of weight and feed overwrote the default values. The main output from the calculator was lifetime N excretion per animal, lifetime P excretion per animal and lifetime GHG emissions per animal as CO2 equivalents.

The eECO FCE Hub
The eECO FCE hub was built from the outputs of the ECO FCE project as a whole. It represents a web based interface (on the ECO FCE project website) where stakeholders can chose to mine research information in the broad areas of feed, gut and genetics for either pigs or broiler chickens. It is aimed to be a user-friendly tool where a range of stakeholders can gain information quickly about the key findings of ECO FCE.
Potential Impact:
ECO-FCE has generated a significant amount of data which has been communicated to many different audiences. The attached table (“Meetings where ECOFCE material has been presented) highlights some of the scientific, technical and outreach meetings where ECO-FCE presentations were made by consortium members during the course of the project. This includes scientific presentations made at 30 separate events in 13 countries over three continents. In addition, it includes workshop events targeted at the agri-food sector, some of which have been organised by the ECO-FCE team, and events targeted at younger stakeholders. The issues of food security and environmental protection are extremely relevant to our youngest citizens, and our outreach events have specifically aimed at providing education in this respect. This includes developing our ECO-FCE exhibition which has been hosted at three science museums across Europe to date, and developing local children’s education events in Poland and Northern Ireland. These ECO-FCE dissemination activities have been accompanied by specially developed booklets containing key information for different types of stakeholder.

The ECO-FCE project team has also been extremely proactive in producing scientific publications. To date 13 peer-reviewed scientific papers have been published or are in press in high impact scientific journals, and a significant number of additional papers have either been submitted to be published or are in the final stages of preparation. In addition, a total of 57 abstracts from scientific conferences have been published to date, and 8 scientific summaries have been published as part of proceedings of agri-food ‘industry’ events. A full overview of scientific publications from the project is available in Deliverable 6.7.

In terms of other dissemination activities, our partners have been proactive in producing press releases, and we have also produced a number of videos from this project (including an ECO-FCE documentary movie, project overviews by the co-ordinator and deputy co-ordinators, presentations at conferences). All of these dissemination activities are available to view on our website ( In addition the main dissemination activities and approaches from ECO-FCE are summarised in the attached document (“ECO-FCE Use and dissemination of foreground”).

The potential impact and exploitation of results relating to specific work packages is listed below:

Work package 1
The main results from WP 1 can be summarised as:
1. Identification of novel feed additives and ingredients to beneficially affect pig and broiler chicken feed conversion efficiency (FCE) and ecological footprint.

In terms of impact, this knowledge has the potential to be exploited by both researchers and feed technologists and could be used to develop new sustainable feed solutions in the future. The feed additives and ingredients identified could also open new revenues of income for some countries due to the nature of their source. On the other hand, the use of these novel additives and ingredients could help the sustainability of monogastric production in countries where food security is uncertain.

2. Establishment of a mega eECO FCE warehouse of information to conduct mini meta analysis.

In terms of impact, this warehouse contains a vast array of trials, results and experimental conditions from which many student projects could emerge, and/or which could help feed technologists to understand the scope of work already completed in specific areas. As such, the future exploitable potential of this database is huge across a range of stakeholders. It is likely it will be mainly utilised by the consortium academic partners to enable the completion of student projects and to identify areas of research requiring further investigation, or those that have been sufficiently explored. Therefore, it could help researchers and academics to build evidence for project proposals or advice to other stakeholders.

A key aspect which is highly exploitable from this section of work is the future potential to build an automated method for real-time capture of information from peer-reviewed publications. The development and quality control steps designed to manage the eECO FCE warehouse have, in themselves, provided researchers with the knowledge of how to expand and automate such an electronic resource going forward. The leader of WP 1 (AFBI) is keen to pursue this methodology with publishing houses to enable the rapid and effective use of information reported in peer-reviewed papers.

3. The lack of robust research in the area of nutrition/genotype/environmental interactions. This supports policy decisions to fund research in this area going forward.

This information is highly valuable to policy makers who require confidence to invest research funding in areas which have much potential to lead to further improvements in food security. As the European agri-food industry strives for increased precision, the ability to feed individual animals of different genotypes represents a key area where knowledge is required to achieve ultimate ‘precision nutrition’.

Work package 2
The use of precision feeding to improve nutrient efficiency in pigs has great potential to reduce the environmental impact of pig farming, mainly due to potentially reducing nitrogen excretion. Excesses or deficiencies in protein supply result in imbalances, thus, precisely matching supply to requirements should minimise N excretion. However, the approach taken to improve precision (i.e. homogenising pigs by sorting them in 3 size groups), appeared not to be sufficient to achieve substantial improvements. Future research should focus on the delivery of tailored feed to a wider range of animal sizes (and ideally to the individual level).

Liquid feeding has been proven to offer different possibilities (e.g. to be used in combination with enzymes, prolonged soaking, fermentation, etc.) for the improvement of nutrient utilisation of feeds for growing-finishing pigs. Particularly large improvements were observed for the efficiency of use of phosphorous, thus reducing its associated environmental impact. Besides, the use of liquid feeding may also have economic benefits associated with improved feed efficiency as well as improving the potential use of non-conventional feed ingredients.

Similarly, phosphorous nutritional conditioning has been shown to improve P utilisation in broiler chickens. This is a promising strategy to reduce the environmental impact of P excretion, and also has economic benefit associated. However, additional work is still needed to achieve the full potential of this approach.

A significant socio- economic impact from the findings of the ingredient and enzyme work was that animal performance was not significantly detrimentally affected when the diet contained high levels of rapeseed meal and wheat distillers grains with solubles compared with when soya bean based diets were offered. The levels used in this work were as high as possible and therefore this work supports the use of such home grown protein sources for the European pig and poultry industries. The ability of the European industry to use home grown protein sources is not only an economic advantage but also an environmental advantage in terms of reducing carbon footprint.

Research within ECO-FCE found a very inconsistent effect of enzyme addition. Among the promising combinations found there were: protease in soya bean meal based diets for poultry and in field-bean based diets for pigs, alpha-galactosidase in rapeseed meal-based diets for pigs, and beta- glucanase –xylanase in extruded barley based diets for poultry. The European pig and poultry industries use significant amounts of enzymes in diets. On one hand significant savings may be possible if the use of enzymes was more targeted, and on the other hand this work highlights the need for more independent research in this area to inform producers and feed formulators on the best use of enzymes for different raw ingredients.

It can be concluded that saponins and essential oils have great potential to reduce the ammonia emissions from fattening pigs and to improve the feed efficiency. Spices and also some essential oils are more suitable in weaning piglets as they are more effective on feed intake. The combination of saponins and essential oils with spices and short and medium chain fatty acids was not always positive and a final conclusion on these complex applications cannot be drawn. Research with broilers shows a clear benefit in broiler performance, i.e. daily gain and feed efficiency, when diets are supplemented with a mixture of essential oils and saponins. The benefits of spices are not conclusive. A combination of essential oils and saponins with other additives, such as antioxidant herbal extracts or short and medium chain fatty acids, did not result in better performance but should potentially be evaluated under challenge conditions.

The piglet perinatal nutrition work (focussing on L-arginine and L-carnitine) showed that it is possible to have a high survival rate of low weight piglets from large litters with the use of rescue decks. This can be an important measure to reduce piglet mortality, which has both animal welfare and economic impacts. In addition, post-weaning growth performance of low birth weight pigs artificially reared was similar to that of low birth weight pigs reared by the sow. However, further work is still required to improve the nutrient composition of milk replacers and to improve the efficiency of low birth weight piglets from large litters.

Work package 3
ECO-FCE research on relationships between the feed use efficiency of chickens and gut related factors (microbiome, gut size, structure and function), yielded several exploitable results. The most exploitable observations regarding similar feed-efficiency related profiles were serum metabolites and the bacterial phylotypes which positively correlated with feed efficiency. Especially, the identified serum metabolites may be exploited as biomarkers for feed efficiency in chicken production. Likewise, based on ECO-FCE findings, further research is warranted to explore the potential of the Dorea-phylotype as a probiotic. Results obtained were presented at eight different international scientific conferences. Moreover, results on chicken’s feed efficiency-related differences were published in or submitted to journal with peer-review and summarized in a press release. In total, one paper was published and three manuscripts were submitted from this work prior to January 31st, 2017.

From the pig work with a similar objective, one ECO-FCE study showed that microbes associated with a leaner and healthier host (e.g. Christensenellaceae, Oscillibacter, Cellulosilyticum) were found to be enriched in low RFI (more feed efficient) pigs, while Nocardiaceae (Rhodococcus) were less abundant in the ileum of low RFI pigs. Therefore, differences, albeit relatively subtle, within the intestinal microbiota of low RFI pigs were found compared to their high RFI counterparts, demonstrating a possible link between the intestinal microbiota and FE in pigs. In the second pig study several FE-associated enterotypes were common across batches and geographic location and related to a potentially “healthier” and metabolically more capable microbiota. These taxa, as well as those mentioned in the first study, could therefore be employed as biomarkers for FE and may merit consideration for use as probiotics or targeting by dietary means as a strategy for improving FE in pigs in the future.

When we examined existing and new feed efficiency metrics within this work package we found that each metric measures a different aspect of efficiency, although improvement in one of the traits would, on average, lead to improvements in the others. These results can be used by stakeholders to decide on which metric should be used for feed efficiency for specific purposes. These results were presented at eleven conferences. Three manuscripts relating to the pig work from this work package have been submitted to high impact factor peer reviewed journals.

This work package also involved intervention studies at Vetmeduni (Austria) and Teagasc (Ireland) designed to promote optimal gut microbiota for animal performance and efficiency. From the intervention strategies tested in chickens post-hatch at Vetmeduni, restrictive feeding was the more promising approach which should be followed-up in future research work. Specific changes in the bacterial community in ileal and caecal digesta caused by the restrictive feeding, such as increased Lactobacillaceae and reduced Turicibacteraceae and Enterobacteriaceae, may be targeted in future studies either by diet formulation, or, in case of the Lactobacillaceae, to optimize probiotic applications in combination with restricted feeding. Further optimization of probiotic applications may not only improve chicken’s feed efficiency but may also help in reducing the therapeutic use of antibiotics. From this respect, although the administration of a faecal inoculum and function proved to be less useful to shape the chicken’s gut microbiome and structure to promote feed efficiency, this approach, as well as the restrictive feeding, should be further investigated to determine effects on gut health. The fact that both posthatch interventions had only small effects on gut size, structure and integrity demands further research to improve our current understanding about the molecular mucosal networking. Results from this work will be presented at the BOKU Symposium in April 2017 and the preparation of manuscripts about the nutritional intervention strategies post-hatch for publication in peer-reviewed journals is underway.

The faecal microbiota transplant (FMT) work in pigs, although conducted to manipulate the intestinal microbiota in order to improve FE, had a negative impact on lifetime pig growth and associated physiological parameters. The results from the two experiments conducted suggest a number of things, including (1) that care must be taken that FMT is conducted at the appropriate stage in the animal’s life, (2) that it is possible to influence the microbiome in the pig so that growth of the animal is affected, (3) that microbiota that negatively affect pig growth can be identified from these studies, and (4) that a more targeted approach to microbiota manipulation as a means of modulating pig growth and feed efficiency is warranted. These results were presented at one conference so far, are submitted for another and will be presented at at least three others over the next 12 months. Three manuscripts relating to this pig work have been submitted to high impact factor peer reviewed journals.

In relation to other tasks from this work package, synbiotic formulations for efficient feed conversion in chicken were developed at UTP together with IBB (both partner organisations in Poland). Both formulations may have impact on animal welfare and health, reducing farm costs. In ovo work and validation of synbiotics provide natural solutions for poultry production (hatcheries) to stimulate a more beneficial microbiota and modulate the immune system of an animal before hatch. EU policy makers might be willing to consider promoting synbiotics as alternatives to veterinary medical preparations (including antibiotics) that are sometimes used at excessively high levels in animal production.

The in ovo results will be further exploited in industrial demonstration works, including synbiotic challenge trials, to further optimize the in ovo technology and prepare it for implementation. The established of an in vitro synbiotic validation protocol will serve to formulate new powerful synbiotics. Results on strain characterization and their persistence in chicken gastrointestinal track were presented at the 11th International Symposium on Lactic Acid Bacteria in The Netherlands. Moreover, a manuscript has been prepared from this work and will be submitted to Applied Microbiology and Biotechnology. Dissemination of remaining research relating to this in-ovo research involves peer-reviewed publications, demonstrating the in ovo technology to industrial stakeholders (presentations in productions sites, publications in trade press), publications in the daily press, interviews, workshops and education events.

Work package 4
The overall project has enormous positive environmental impact. However there was potential for the resultant meat from more efficient animals and systems to be of greatly reduced quality. Willingness-to-pay studies indicate that consumers are prepared to pay a premium for a high quality product but are not willing to purchase defective and impaired quality product. The outputs of this project will go some way to offering reassurance to pig and chicken producers and retailers that meat from more efficient animals is of acceptable quality. The drive towards efficiency will permit high quality pig meat and chicken to be offered to consumers throughout Europe.

Under these prerequisites, the increase in feed efficiency of livestock species can be beneficial for the competitiveness of European agri-food production and the global availability of added-value agricultural goods which rely on the efficient and sustainable usage of resources. This includes animal feed in general, but certain irreplaceable micronutrients such as phosphorus (P) in particular. Therefore, the assessed strategies to reduce P intake via nutritional conditioning and animal-specific adaptive responses provide first steps towards redefining animal- and age-specific requirements for micronutrients in diets. The implementation of such measures might have implications on the environment via the reduction of losses and wastes, on agri-food production by reducing of feed costs, and on society via an increased acceptability of animal production systems due to a reduced ecological footprint while sustaining product quality and safety.

Nevertheless, molecular analyses suggests that the improvement in efficiency is not endless, particularly considering obvious trade-offs between efficiency and resilience traits, especially under diversified conditions. Therefore, molecular biomarkers that are indicative for the partitioning and utilisation of nutrients and the animal’s metabolic state, as well as for its responsiveness to different feeding regimes, were proposed within ECO-FCE. This comprises hub molecules of macronutrient metabolism which need to be proven as proxies for an early method of determination of favoured metabolic routes. Moreover, genome-wide analyses shed light in the genetics of FE traits, thus providing relevant candidate genes that might be considered for the implementation in animal breeding or building the bases to improve FE traits via new genome editing approaches. In pigs, this information was generated and discussed in close cooperation between FBN (Partner 12), Hermitage (Partner 5) and Aarhus (Partner 11) in order to implement genomic information in the breeding programme of Hermitage Genetics. Accordingly, Hermitage started to build a phenotyped resource population of different pure and commercial breeds by routine ear tag sampling. Regarding the SNP-chips, Hermitage assesses the ordering and utilization of a custom 3K chip based on the selected SNPs. However, under consideration of the price trend for commercial high density SNP chips, an available higher density chip (GGP-HD 80K) was finally used for the implementation of genomic selection for FE.

The project has furthermore improved research links between European researchers and their US counterparts, through a Teagasc-funded 3 month study visit to the Agricultural Research Service, Nebraska, wherein genetic markers identified as significant in European pigs, were validated in US pig populations. The information generated on the eating quality of pigs and broilers in this project will be an excellent resource for industry.

Information regarding the effects on feed efficiency via nutritional modulation, nutritional conditioning and genetic strategies, as well as their consequences for meat eating quality, has been presented widely over the course of the project. The scientific community was addressed through international and local conference presentations, including the International Congress of Meat Science and Technology 2015, the World Poultry Congress 2016, the Congress of European Association of Animal Production (EAAP) in 2015 and 2016, the Genetic Days 2016, and the Polish Congress of Genetics 2016. Moreover, representatives presented ECO-FCE results at industry-focused events such as the Teagasc Gateways events, the EuroTier 2016, and the Polish Congress of Genetics in two poultry symposia entitled “Science for practice and practice for science”. Specifically, a workshop on “Developing tools to maximise feed efficiency in monogastrics” was organized in Warsaw, as a satellite meeting to the 66th congress of the EAAP. It should also be highlighted that information from this and other workpackages was also presented at several seminars at the partner’s institutes/universities. Final project outcomes will be also presented at upcoming conferences like EAAP and ISAG 2017, the 10th European Symposium on Poultry Genetics, and the 19th International Conference on Farm Animal Genetics in Spain.

The published and projected scientific articles describe the application of different holistic approaches to datasets of pig and broiler chicken populations in order to contribute to the elucidation of the genetic architecture of feed efficiency traits. The obtained results for both chicken and pig populations were submitted for publication in peer-reviewed scientific journals (seven original papers). Moreover, one review on the genetic background of FE in broilers is submitted. Regarding the meat-quality data and the transcriptomic analyses, several peer-review publications are currently in preparation and intended to be submitted in 2017. A meat quality database has been established which is accessible by industry partners. A Teagasc Technology Update of the database will be prepared in regular terms.

Different kinds of press articles were released in the countries of Work Package 4 partners to disseminate the research activities on host genetic factors at the regional and national level. Therefore, press articles focused on the genetic architecture of feed efficiency traits and the molecular foundation of divergent feed use efficiency. The general ideas of the ECO-FCE project were published in most important poultry magazine in Poland - “Polskie Drobiarstwo”, in a local newspaper in Germany, and the ‘Leibniz Nordost’ which is the journal of the Leibniz Institutes located in Mecklenburg-West Pomerania. Moreover, the main important genetic results (from a practical point of view) from chicken and pig data were published in two magazines by Poznsn University (Poland): “Polskie Drobiarstwo” and “Trzoda chlewna”, respectively.

Work package 5
The mission within this work package was to develop statistical methods/tools and approaches for improved genetic decisions by farmers/breeders. The advancement in the statistical methods achieved in ECO-FCE can improve genetic progress in livestock populations for economically and environmentally important traits across Europe. This will result in the production of efficient and environmentally friendly animals. The results will have impact on the farming/policy making decisions in farm and European level.

In terms of detail relating to potential impact, the guidelines to use genetic residual feed intake as a measure of feed efficiency has the potential to revolutionize the way animal breeding has been conducted. This may mean a new era of breeding by more accurately selecting for improved feed efficiency. Utilising the longitudinal feed efficiency approach also allows selection and improvement of the growth curve of animals; an approach that has not been able to be implemented due to lack of statistical methods. The Horizontal method shows the implementation of this method in the breeding programmes to be feasible and to provide added value in a cost-effective manner. This method allows the utilisation of longitudinal feed intake data with single recorded production traits to obtain longitudinal genetic residual feed intake; with impact of updating the way selection of elite animals has been achieved. This results in production of more efficient and environmentally friendly animals in Europe.

The novel genomic feature model impacts largely on the accuracy of finding important regions associated with traits of interest, as well as leading to substantial improvement in identifying the elite animals for economically important traits. In addition, the horizontal model and the genomic feature model can be implemented across species and on large numbers of traits.

The impact of the project can be also be seen by the large numbers of scientific manuscripts that have been produced throughout the project. In addition, this project has led to further collaborations between ECO-FCE industry and academic partners in terms of a new project to further investigate the implementation of integrative models in genetic selection programmes for broiler chickens. This project also provided the genomic reference population and the genomic models necessary to obtain the genomic breeding values for animals in the Hermitage MaxGro pig population.

The eECO FCE hub was developed in Work Package 5 and provides a great resource to ensure the long term impact of the project as a whole, since it allows stakeholders to access key project findings in an easy-to-access and succinct manner. The 3-in-1 ecological calculators allow researchers and feed technologists to make simple environmental impact assessments on the nutrition work they conduct, without having to conducted detailed analysis. This knowledge will help inform researchers and feed technologists about the holistic impact of their work and future strategies to investigate, which both improve animal performance as well as reduce the ecological footprint of monogastric production.

These ecological calculators were used to consider the ecological impact of some of the experimental treatments assessed in ECO-FCE. A total of four pig related experiments and four broiler related experiments conducted in WP 2 were used in this exercise. In each of these experiments the N, P and GHG emissions were compared between the control treatments and the treatment which either demonstrated statistically the best pig performance, and in the case where no statistical difference was found, the treatment which led to the largest effect. The results from this exercise are presented in Tables 1 and 2 of the attached document (“Ecological calculator tables”) demonstrate the relatively large effect that dietary treatments can have on the ecological footprint of pig and broiler chicken production.
List of Websites: