R. Roehe

Genetic association between feed intake and feed intake behaviour at different stages of growth of group-housed boars

Abstract Voluntary feed intake, feed intake pattern and performance traits were recorded on 3188 group housed boars of Landrace and Large White tested from day 100 to day 170. Measurements of feed intake and feed intake behaviour were obtained by electronic feed dispensers (ACEMO) under ad libitum conditions. Heritabilities of feed intake in periods 1 to 5 and over the entire test period were estimated to be 0.16, 0.24, 0.30, 0.27, 0.26, and 0.22, respectively. The maximum heritability of daily feed intake in the third time period on test corresponds to about 130 days of age at about 85 kg of weight. Daily feed intake in this period showed a high genetic correlation with the average feed intake over the whole performance test (rg = 0.91). Estimates of genetic correlations between daily feed intake and feed conversion ratio, residual feed intake, average daily gain or backfat thickness were 0.04, 0.97, 0.62 and 0.42, respectively. Boars feed intake activities decreased over time while time per day in the feeder was almost constant. Traits of feed intake behaviour as feeding rate, feed intake/visit, number of visits, time per visit, and time per day in the feeder showed high heritabilities of 0.44, 0.51, 0.43, 0.42, and 0.43, respectively, but genetic correlations with performance and carcass traits were generally low. One exception was the behavioural trait time per day in the feeder with its moderate genetic correlations to daily feed intake (rg = 0.44) and average daily gain on test (rg = 0.32). Selection for lean growth will improve lean growth feed efficiency as indicated by the negative correlation of −0.47. But more efficient may be the use of the component traits, lean content, daily feed intake and daily gain in particular of the most informative test period.

Associations of DNA markers with meat quality traits in pigs with emphasis on drip loss

Phenotypic information on 1155 market pigs for several pig meat quality traits, was collected. Genotypes on 12 DNA markers, including RYR1 and PRKAG3 I199V, were also obtained on all pigs to investigate the relationship between genetic markers and meat quality. The RYR1 gene had the highest impact on meat quality, however, several other markers showed significant effects on one or more traits. Animals heterozygous at the RYR1 locus were significantly inferior in almost all meat quality traits, except ultimate pH value, initial conductivity and redness of the meat. Drip loss from case-ready meat (measured from 1 to 7 days post-mortem) was 43% higher for heterozygotes than animals of the stress resistant genotype. The homozygous genotype II at position I199V of the PRKAG3 locus also resulted in less drip loss than genotypes IV and VV, regardless of the method and time of measurement. Furthermore, the favourable genotype related to higher ultimate pH and darker meat. Both loci significantly affected the intercept, linear and quadratic terms of fitted drip loss development curves. The favourable genotypes showed a lower drip loss after one day of measurement and a slower increase and a more linear development over time. Whilst the RYR1 and PRKAG3 markers influenced numerous meat quality traits, some of the other markers were also found to have significant effects on one or two meat quality traits. Markers at MC4R and HMGA1 loci significantly affected drip loss, whereas LDHA, CAST (Hpy188I) and ATP2A1 influenced pH value. In addition, the marker ATP2A1 was associated with variation in intramuscular fat content in M. longissimus dorsi. GLUT4 affected temperature 45min post-mortem and several markers (MC4R, LDHA, GLUT4, HMGA1, CAST (Hpy188I and PvuII)) influenced one or two of the different colour measurements. The markers at MC4R, CKM, AGRP, PRKAG3, and HMGA1 loci were tested for their interactions with RYR1 regarding drip loss. Only AGRP showed a significant interaction, but this was based on only a few animals with the homozygous genotype for one allele. Our results suggest that genetic markers provide a useful tool to improve meat quality in pigs independently from RYR1, especially the mutation I199V in the PRKAG3 gene.

Comparison of different methods for determination of drip loss and their relationships to meat quality and carcass characteristics in pigs

Samples of the M. longissimus dorsi of 776 pigs from three commercial lines were used to compare two methods for measuring drip loss, referred to as the EZ-DripLoss and bag methods. Furthermore, relationships between drip loss and other meat quality and carcass traits were analysed. The bag method used a slice of M. longissimus dorsi of approximately 120 g hung in an airtight container whereas the EZ-DripLoss method used two samples of approximately 10 g placed in drip loss containers. In the bag method, samples taken at 24 h post-mortem were measured 24 and 48 h after sampling and average drip loss was 1.94% and 3.33% at 24 and 48 h, respectively. Correlation between these consecutive measurements was high (r=0.98). Using the EZ-DripLoss method, drip loss of samples taken at 24 h post-mortem was measured 48 h after sampling and showed an average value of 4.97%. Correlation between the drip loss obtained using EZ-DripLoss and bag methods was high (r=0.86). Relationships between drip loss and other meat quality traits were similar for both methods. Highest correlations were observed between drip loss and pH(45) (r=-0.52 and -0.48 using EZ-DripLoss method(48) and bag method(48), respectively) and the lowest to Minolta a (∗) value (r=0.11 and 0.09, respectively). Correlations among several carcass traits, such as lean content, and drip loss were low or not significant. Low associations between loin eye area (cm(2)) and drip loss were obtained regardless of the method used to determine drip loss (r=0.21 and 0.18 using EZ-DripLoss method(48) and bag method(48), respectively). For routine measurements, the EZ-DripLoss method is recommended because it showed a high correlation with the bag method but is easier to perform and is more standardised.

Developments of carcass cuts, organs, body tissues and chemical body composition during growth of pigs

Abstract A serial slaughter trial was carried out to examine the developmental change of physical and chemical body compositionin pigs highly selected for lean content. A total of 48 pigs (17 females and 31 castrated males) were serially slaughteredand chemically analysed. Eight pigs were slaughtered at 20, 30, 60, 90, 120 and 140kg live weight, (LW) respectively.The carcass was chilled and the left carcass side was dissected into the primal carcass cuts ham, loin, shoulder, bellyand neck. Each primal carcass cut was further dissected into lean tissue, bones and rind. Additionally, the physical andchemical body composition was obtained for the total empty body as well as for the three fractions soft tissue, bones andviscera. Viscera included the organs, blood, empty intestinal tract and leaf fat. The relationship between physical orchemical body composition and empty body weight (EBWT) at slaughter was assessed using allometric equations (log 10 y ¼ log 10 a þ b log 10 EBWT). Dressing percentage increased from 69·4 to 85·2% at 20 to 120kg and then decreased to83·1% at 140kg LW, whereas percentage of soft tissue, bones and viscera changed from 23·5 to 33·0%, 10·1 to 6·3%and 14·7 to 10·3%, respectively, during the entire growth period. Substantial changes in proportional weights of carcasscuts on the left carcass side were obtained for loin (10·5 to 17·5%) and belly (11·3 to 13·8%) during growth from 20 to140kg. Soft tissue fraction showed an allometric coefficient above 1 (b ¼ 1·14) reflecting higher growth rate in relation tothe total empty body. The coefficients for the fractions bones and viscera were substantially below 1 with b ¼ 0·77 and0·79, respectively, indicating substantial lower growth relative to growth of the total empty body. Lean tissue allometricgrowth rate of different primal cuts ranged from b ¼ 1·02 (neck) to 1·28 (belly), whereas rates of components associatedwith fat tissue growth rate ranged from b ¼ 0·62 (rind of belly) to 1·79 (backfat). For organs, allometric growth rate rangedfrom b ¼ 0·61 (liver) to 0·90 (spleen). For the entire empty body, allometric accretion rate was 1·01, 1·75, 1·02 and 0·85for protein, lipid, ash and water, respectively. Extreme increase in lipid deposition was obtained during growth from 120 to140kg growth. This was strongly associated with an increase in backfat and leaf fat in this period. Interestingly, breedsselected for high leanness such as Pie´train sired progeny showed an extreme increase in lipid accretion at a range of LWfrom 120 to 140kg, which indicates that selection has only postponed the lipid deposition to an higher weight comparedwith the normally used final weight of 100kg on the performance test. The estimates obtained for allometric growth ratesof primal carcass cuts, body tissue and chemical body composition can be used to predict changes in weight of carcasscuts, determine selection goals concerning lean tissue growth, food intake capacity, etc. and generally as inputparameters for pig growth models that can be used to improve the efficiency of the entire pig production system for pigshighly selected for lean content.Keywords: allometry, body composition, carcass composition, growth, organs, pigs.

Association between body composition of growing pigs determined by magnetic resonance imaging, deuterium dilution technique, and chemical analysis.

Development of body composition of 440 growing pigs from a three generation full-sib design to identify quantitative trait loci (QTL) was determined by three different methods. Firstly, the non-invasive method deuterium dilution technique (DT), was applied to all pigs in the experiment at six weights 20, 30, 60, 90, 120 and 140kg. Secondly, at each weight class, eight pigs were slaughtered and their entire body chemically analysed (CA). Thirdly, magnetic resonance imaging (MRI) was applied on 16 live pigs at different weights. For the entire empty body (without content of the gastrointestinal tract and bladder), allometric prediction equations to predict body composition from empty body water content measured by DT were derived from chemically analysed serial slaughtered pigs. These equations showed high correlations of 0.92, 0.90 and 0.85 for the contents of body water, fat-free substance as well as protein in fat-free substance, respectively. For the soft tissue (empty body without bones and viscera), allometric prediction equation of body composition based on DT and CA showed correlations of 0.91, 0.88 and 0.82 for water content, fat-free substance, and protein content of fat free substance, respectively. Fat tissue content, fat tissue mass, and lean tissue mass measured by MRI showed allometric relationships to lipid content, lipid mass, and protein mass determined by DT with correlations of 0.98, 0.87, and 0.98, respectively. Lean (measured by MRI) and protein (determined by DT) content of soft tissue was best fitted by a linear-quadratic polynomial and resulted in a correlation of 0.86. Allometric coefficients for change of percentages of chemical components, water (b=-0.036) and protein (b=0.106) in fat-free substance of empty body during growth were similar to those in the literature indicating the consistency of accretion rates of chemical components of the fat-free substance in different studies. Means for protein- and lipid-deposition rates (determined by DT) as well as lean tissue- and fat tissue-deposition rates (measured by MRI) ranged from 95 to 154, 147 to 328, 373 to 420 and 129 to 254g in the different weight ranges. Variation between animals in protein (lean tissue) and lipid (fat tissue) deposition rate was large which can be exploited in order to identify QTL of these traits.

Comparison of linear and nonlinear functions and covariance structures to estimate feed intake pattern in growing pigs

Abstract The objective of this study was to find the best function and covariance structure to estimate feed intake pattern in growing pigs for breeding purposes in order to optimise the feed intake curve corresponding to lean growth rate. Daily feed intake from 81 group-housed pigs was recorded using electronic feeders. The animals were tested during 120 days on average, from 30 to 119 kg. Polynomials, Kanis, yield–density, segmented and sigmoidal functions showed high similarity in goodness-of-fit. For selection on early feed intake, linear-segmented, logistic and Richards functions resulted in the most usable estimates within the test period. As shown by simulation, parameters of logistic function resulted in the lowest bias. Covariance among residuals of subsequent daily feed intakes was accounted for using the structures variance components, compound symmetry (CS), first-order autoregressive AR[1], first-order autoregressive moving-average ARMA[1,1], heterogeneous CS, heterogeneous AR[1] and the power-of-the-mean variance model. Correlation structure ARMA[1,1] resulted in the best fit of the data, with estimates for autoregressive parameter ρ from 0.90 to 0.97 and for moving average parameter λ from 0.35 to 0.51. The power-of-the-mean variance model was a good characterisation of variance heterogeneity and the final estimated power was 2.008 with standard error 0.3245. Based on these results, linear-segmented and logistic functions were the most parsimonious functions to characterise feed intake from which selection criteria can be derived, such as the age at which the feed intake plateau or the age at which the maximum increment in feed intake per day is reached in order to change feed intake curve corresponding to lean growth curve.

Genetic associations between observed feed intake measurements during growth, feed intake curve parameters and growing–finishing performances of central tested boars

Abstract Optimization of growing–finishing performance can be reached by increasing feed intake capacity of pigs at an early stage of growth and by limiting feed intake at the end of the finishing period. Daily feed intake recorded by electronic feeder and other growth performances were obtained from 5601 group-penned nucleus boars of two dam lines on a central test station. A multiple trait animal model was used which considers the observed feed intake at each test week as a different trait. Estimated heritabilities for feed intake were 0.12, 0.23, 0.29, 0.39, 0.32 and 0.39 for the first, third, fifth, seventh, and ninth test week, and entire test period, respectively. A linear regression and a linear-quadratic regression were fitted for daily feed intake as functions of time. Heritabilities of the function parameters ranged between 0.11 and 0.32. Average estimated feed intake and average change in daily feed intake (slope) were obtained for five periods (12 days) based on linear-quadratic regression. Heritabilities for average estimated feed intake changed from 0.17, 0.32, 0.37, 0.43 to 0.41 and the slope of feed intake from 0.18, 0.21, 0.33, 0.27 to 0.22 with time on test. Genetic correlations of observed feed intake among different periods were substantially different from one (0.02–0.73) which indicates that feed intake was differently genetically determined during growth. In order to account for the different genetic relationships of feed intake during performance test and their influence on other performance traits, a multiple trait animal model using average estimated feed intake of each period, slope of the first three periods, and all genetic associations among these traits, were recommended to improve feed efficiency and lean growth.

Allometric association between in vivo estimation of body composition during growth using deuterium dilution technique and chemical analysis of serial slaughtered pigs

The objective of this study was to develop accurate mathematical-statistical functions to estimate body composition of live pigs between 20 and 140 kg weight from total body water (TBWA) determined by the deuterium dilution technique. Chemical body compositions during the growth period are essential input parameters for biological pig growth models, which are used to estimated the nutrient requirements, improve the entire production system, determine optimal slaughter weight, optimize selection for food intake, etc. In the present study, 48 pigs (17 female and 31 castrated males) were used in an experimental station to obtain protein, lipid, ash and water content at 20, 30, 60, 90, 120 and 140 kg live weight. At each target weight, body water of the animals was determined by the deuterium dilution technique. Eight pigs of each live-weight group were slaughtered and chemically analysed. Water content of the empty body decreased from 74 to 53%, whereas lipid content rose from 7 to 30%. Between 20 and 30 kg body weight, protein content increased from 16 to 17% and thereafter decreased to 16%. Ash content was constant at 3%. To estimate body composition of the remaining animals from TBWA (%) determined by deuterium dilution technique, two sets of exponential prediction functions were used to describe the relationship between chemically analysed body components and TBWA (%). The first set of prediction functions fitted one intercept for the entire growth period and the second set of prediction functions fitted a different intercept for each weight class. Correlation coefficients between estimated and chemically determined empty body water, lipid, protein and ash for the first set of functions were 0·93, 0·86, 0·83 and 0·65, respectively. The second set of prediction functions showed higher accuracy (2 to 10%), but had the disadvantage of non-continuous estimates over the entire growth period. In contrast, by using the first set of prediction functions, a continuous accurate estimation of body composition of live pigs was obtained over a large range of growth (20 to 140 kg) based on deuterium dilution space.

Random regression to model genetically the longitudinal data of daily feed intake in growing pigs

Abstract The objective of this study was the development of a genetic statistical method for longitudinal data of daily feed intake in growing pigs using random regression models (RRM) in order to change the pattern of feed intake by selection. Besides a quadratic RRM for additive genetic effects, different combinations of covariance structures for permanent environmental effects using RRM and for temporary residual effects were fitted to individual information from 5245 boars (2938 of line 3, 2307 of line 4) in order to obtain the best fitting model for daily feed intake recorded during a test period of 10 weeks. Based on Akaike’s Information Criterion, the parsimonious model included a constant diagonal covariance structure for the permanent environmental effects and a heterogeneous autoregressive-moving-average structure of order (1,1) for temporary residual effects. Estimates of the autocorrelation were 0.80 (line 3) and 0.81 (line 4) and estimates of the moving-average components were 0.15 (line 3) and 0.11 (line 4). Estimates of heritabilities were low and increased from 0.02 to 0.06 during the test, with the highest estimates in the seventh week on test. The low heritabilities were only due to a higher residual variance when using individual daily records of feed intake, whereas additive genetic variances (0.01 to 0.04 kg2) were similar to estimates using average records of feed intake. Genetic correlations between intercept and linear regression coefficients were positive within 0.57 and 0.55, those between intercept and quadratic regression coefficients were negative within −0.77 and −0.89, and those between linear and quadratic coefficients were within −0.48 and −0.61 for lines 3 and 4, respectively, indicating the opportunity for changing the feed intake pattern. Additionally, eigenfunctions obtained from the additive genetic (co)variance matrix indicated that a genetic change in feed intake pattern is achievable, e.g. the second major eigenfunction, which explained about 10% of the additive genetic variation, allowed selection for a high increase in feed intake at the beginning of the test period and for reduced feed intake at the end of the test period in order to improve feed efficiency.

The influence of feeding behaviour on feed intake curve parameters and performance traits of station-tested boars

Abstract The use of feed intake behaviour traits, obtained from electronic feeding stations, and feed intake curve parameters to genetically improve performance traits was analysed. Daily feed intake and feed intake behaviour of 5601 group-penned boars of two dam lines were recorded by electronic feeders in weeks 1, 3, 5, 7 and 9 during 10 weeks (100–170 d) on performance test. Additionally, performance test traits and parameters of an individually fitted linear-quadratic regression of feed intake on time on test, were available. A multiple trait animal model considering observed feed intake and feed intake behaviour at each test week as a different trait was used. Estimated heritabilities for feed intake and feed intake behaviour of the entire test period were 0.39, 0.46, 0.34, 0.44, 0.44 and 0.41 for daily feed intake, time per day, visits per day, time per visit, feed intake per visit, and feed intake rate, respectively. Heritabilities for behavioural traits in each test week were below estimates of entire test and showed lower variation, except for daily feed intake and time per day. Residual standard deviation of feed intake using linear-quadratic regression on time on test showed a moderate heritability of 0.22. Genetic correlations of feed intake behaviour traits indicate that number of visits per day was independent from growth performance, while time per day was genetically associated with average daily gain and daily feed intake (0.31 and 0.41). In contrast, visits per day and related traits such as time per visit and feed intake per visit were genetically correlated with residual standard deviation of linear-quadratic regression (−0.28, 0.31 and 0.39). Feed intake per visit in the seventh test week resulted in the highest genetic correlation (0.45) with residual standard deviation of the feed intake curve. This residual standard deviation and related traits will be of increasing interest for future breeding programs in order to obtain high performances in a wide range of environments.