Effect of selected water temperatures used in Mycoplasma gallisepticum vaccine reconstitution on titer at selected time intervals.

Numerous methods are currently used throughout the poultry industry for the administration of vaccines. Each utilizes water for vaccine reconstitution and/or administration, including two of the three commercially available live Mycoplasma gallisepticum (MG) vaccines. Selected water temperatures were used to reconstitute and/or dilute the three commercially available live MG vaccines. Water temperatures included 4 C, 22 C (room temperature), and 32 C, and titer (color change units) was recorded at four time intervals, at point of reconstitution (time 0), 15, 30, and 60 min postreconstitution of the vaccines (time periods 15, 30, and 60, respectively). Results for F strain MG (FMG) vaccine showed significant decreases in titer from time 0 to time 15 for the 22 C and 32 C water temperatures but no significant decrease for any time period for FMG reconstituted with 4 C water. For 6/85 strain MG no significant difference in titer was noted for any of four time periods within any of the three water temperatures. For ts-11 strain MG a significant decrease was observed in titer at each of the four postdilution time periods when diluted with 32 C water. There was no significant decrease in titer at any time period for ts-11 MG vaccine when diluted with either 4 C or 22 C water.

Stocking density and physiological adaptive responses of broilers.

Three trials were conducted to assess the effects of stocking density on physiological adaptive responses of broilers. Male broilers were reared in floor pens under conditions similar to those used commercially in the United States. Accepted indicators of adaptation to a stressor were measured on d 49 including plasma concentrations of corticosterone, glucose, cholesterol, and total nitrites as an indicator of nitric oxide, as well as heterophil:lymphocyte ratio. In trial 1, calculated stocking densities were 20, 25, 30, 35, 40, 45, 50, and 55 kg of BW/ m2 and in trials 2 and 3, stocking densities were 30, 35, 40, and 45 kg of BW/m2. Stocking densities were calculated based on a final BW of 3.3 kg. Linear trend analyses were used to assess the role of stocking density on each of the physiological parameters. Results indicate that stocking density did not cause physiological adaptive changes indicative of stress.

Comparison of Gompertz and neural network models of broiler growth.

Neural networks offer an alternative to regression analysis for biological growth modeling. Very little research has been conducted to model animal growth using artificial neural networks. Twenty-five male chicks (Ross x Ross 308) were raised in an environmental chamber. Body weights were determined daily and feed and water were provided ad libitum. The birds were fed a starter diet (23% CP and 3,200 kcal of ME/kg) from 0 to 21 d, and a grower diet (20% CP and 3,200 kcal of ME/ kg) from 22 to 70 d. Dead and female birds were not included in the study. Average BW of 18 birds were used as the data points for the growth curve to be modeled. Training data consisted of alternate-day weights starting with the first day. Validation data consisted of BW at all other age periods. Comparison was made between the modeling by the Gompertz nonlinear regression equation and neural network modeling. Neural network models were developed with the Neuroshell Predictor. Accuracy of the models was determined by mean square error (MSE), mean absolute deviation (MAD), mean absolute percentage error (MAPE), and bias. The Gompertz equation was fit for the data. Forecasting error measurements were based on the difference between the model and the observed values. For the training data, the lowest MSE, MAD, MAPE, and bias were noted for the neural-developed neural network. For the validation data, the lowest MSE and MAD were noted with the genetic algorithm-developed neural network. Lowest bias was for the neural-developed network. As measured by bias, the Gompertz equation underestimated the values whereas the neural- and genetic-developed neural networks produced little or no overestimation of the observed BW responses. Past studies have attempted to interpret the biological significance of the estimates of the parameters of an equation. However, it may be more practical to ignore the relevance of parameter estimates and focus on the ability to predict responses.

Stocking density effects on male broilers grown to 1.8 kilograms of body weight.

This study examined the effects of stocking density on live performance, physiological stress level indicators, and processing yields of male broilers grown to 1.8 kg. A total of 3,120 Ross x Ross 708 male chicks was placed into 32 floor pens (5.57 m2/pen). Stocking density treatments were 25, (75 birds/pen), 30 (90 birds/ pen), 35 (105 birds/pen), and 40 (120 birds/pen) kg of BW/m2. The BW gain, feed consumption, and feed conversion were adversely affected with increasing stocking densities by 35 d. Physiological stress indicators (plasma corticosterone, glucose, cholesterol, total nitrites, and heterophil:lymphocyte) were not affected. Litter moisture was higher as stocking density increased, which led to higher footpad lesion scores. In parallel to growth responses, carcass weight was depressed by increasing stocking density, but carcass yield, absolute and relative amounts of abdominal fat, and carcass skin defects were not affected. Increasing stocking density decreased breast fillet weight and its relative yield and breast tender weight, but not breast tender yield. As calculated stocking density increased 5 kg of BW/m2 beyond 25 kg of BW/ m2, final BW and breast fillet weight decreased by 41 and 12 g, respectively. We conclude that increasing stocking density beyond 30 kg of BW/m2 adversely affects growth responses and meat yield of broilers grown to 1.8 kg but does not alter physiological stress indicators.

Stocking density effects on growth performance and processing yields of heavy broilers.

This study examined responses of male broilers during a 49-d production cycle to 4 placement densities in 2 trials. Trials were pooled because no treatment x trial interaction occurred. In each trial, 1,488 male chicks were randomly placed into 32 floor pens to simulate final densities of 30 (37 chicks/pen), 35 (43 chicks/ pen), 40 (50 chicks/pen), and 45 (56 chicks/pen) kg of BW/m2 of floor space based on a projected final BW of 3.29 kg. Growth rate and nutrient utilization were similar (P > or = 0.05) among the treatments from 1 to 32 d of age. From 1 to 49 d, BW gain (P = 0.011) and feed consumption (P = 0.029) were adversely affected by increasing the placement density from 30 to 45 kg of BW/m2 of floor space. The reduction in cumulative BW gain due to placement density can be partially explained by less feed consumption as evidenced by 95.4% of the sums of squares of BW gain being attributable to feed consumption. Litter moisture content (P = 0.025) and foot pad lesion score (P = 0.001) increased linearly with increasing placement density. Upon processing, whole carcass and breast meat yields relative to BW were not affected (P > or = 0.05) as density increased from 30 to 45 kg/m2. The proportion of whole carcasses with scratches, but not tears, on the back and thighs increased (P = 0.021) as density increased. These results indicate that increasing the density beyond 30 kg/m2 elicited some negative effects on live performance of heavy broilers.

A self-propelled, constant-speed spray vaccinator for commercial layer chickens.

Vaccination of commercial layer chickens is labor intensive and often results in poor rates of seroconversion, which, in turn, generally correlate with decreased flock uniformity and performance. Attempts to improve the vaccination process include numerous variations of individual shop-built vaccinators in use by the layer sector of the poultry industry. Each of these vaccinators has limitations that contribute to poor vaccinations. Major problems include the nonuniform speed of the applicator system and pressure fluctuations at the spray nozzles, which contribute to sporadic dispersion of the vaccine as the vaccinator is pushed or carried past the cages. A battery-powered, self-propelled, constant-speed vaccinator was designed and constructed that operates with constant nozzle pressure. In field use, this vaccinator has resulted in both labor savings (reduction of manpower from five to one to vaccinate 75,000 chickens) and time savings (from 45 min to 7.5 min/poultry house) as well as improved vaccination results (higher positive seroconversion rates) against the poultry pathogen Mycoplasma gallisepticum (MG), a bacterium associated with losses of 15.7 eggs/hen over a 45-wk laying period in MG-infected layers as compared with layers maintained free from infection with MG.

A comparison of fitting growth models with a genetic algorithm and nonlinear regression.

A genetic algorithm (GA), an optimization procedure based on the theory of evolution, was compared with nonlinear regression for the ability of the 2 algorithms to fit the coefficients of poultry growth models. It was hypothesized that the nonlinear approach of using GA to define the parameters of growth equations would better fit the growth equations than the use of nonlinear regression. Two sets of growth data from the literature, consisting of male broiler BW grown for 168 and 170 d, were used in the study. The growth data were fit to 2 forms of the logistic model, the Gompertz, the Gompertz-Laird, and the saturated kinetic models using the SAS nonlinear algorithm (NLIN) procedure and a GA. There were no statistical differences for the comparison of the residuals (the difference between observed and predicted BWs) of growth models fit by a GA or nonlinear regression. The plotted residuals for the nonlinear regression and GA-determined growth values confirmed observations of others that the residuals have oscillations resembling sine waves that are not represented by the growth models. It was found that GA could successfully determine the coefficients of growth equations. A disadvantage of slowness in converging to the solution was found for the GA. The advantage of GA over traditional nonlinear regression is that only ranges need be specified for the parameters of the growth equations, whereas estimates of the coefficients need to be determined, and in some programs the derivatives of the growth equations need to be identified. Depending on the goal of the research, solving multivariable complex functions with an algorithm that considers several solutions at the same time in an evolutionary mode can be considered an advantage especially where there is a chance for the solution to converge on a local optimum when a global optimum is desired. It was concluded that the fitting of the growth equations was not so much a problem with the fitting methodology as it is with the form of the equation.

The power of tests for bioequivalence in feed experiments with poultry.

Several studies have compared the feeding of genetically modified (GM) grains and conventional grains to poultry. The general conclusion has been that there were no significant differences detected in the biological performance of the birds (i.e., the grains were bioequivalent). However, the question has been posed whether the experimental designs used in the studies had sufficient statistical power to detect treatment differences. The power of tests can be used to determine the ability of an experimental design to detect treatment differences. The definition of statistical power is the probability of rejecting the null hypothesis when it is false and should be rejected. The complement of statistical power is the Type II error (beta). That is, accepting the null hypothesis that there is no difference in treatments when there is one. A priori power analysis can indicate the probability at which the sampling regimen or experiment can actually detect an effect if a difference exists. Post hoc power analysis indicates the sufficiency or the sample size needed for an experiment that has already been conducted. In the current study, the power of tests for experiments published in the literature where significant and nonsignificant differences were reported between control birds and birds fed new feed grains was examined. With some exceptions, the power of tests is rarely formally considered or mentioned in poultry research. The results of the survey of the literature showed, in general, low power of statistical tests for feeding experiments involving non-GM grains or in those cases when GM and non-GM grains were compared in poultry feeding experiments. These results suggest that care needs to be taken when designing experiments for bioequivalence of grains fed to poultry.

Optimization of phase feeding of starter, grower, and finisher diets for male broilers by mixture experimental design: forty-eight-day production period.

A mixture experiment, a variant of response surface methodology, was designed to determine the proportion of time to feed broiler starter (23% protein), grower (20% protein), and finisher (18% protein) diets to optimize production and processing variables based on a total production time of 48 d. Mixture designs are useful for proportion problems where the components of the experiment (i.e., length of time the diets were fed) add up to a unity (48 d). The experiment was conducted with day-old male Ross x Ross broiler chicks. The birds were placed 50 birds per pen in each of 60 pens. The experimental design was a 10-point augmented simplex-centroid (ASC) design with 6 replicates of each point. Each design point represented the portion(s) of the 48 d that each of the diets was fed. Formulation of the diets was based on NRC standards. At 49 d, each pen of birds was evaluated for production data including BW, feed conversion, and cost of feed consumed. Then, 6 birds were randomly selected from each pen for processing data. Processing variables included live weight, hot carcass weight, dressing percentage, fat pad percentage, and breast yield (pectoralis major and pectoralis minor weights). Production and processing data were fit to simplex regression models. Model terms determined not to be significant (P > 0.05) were removed. The models were found to be statistically adequate for analysis of the response surfaces. A compromise solution was calculated based on optimal constraints designated for the production and processing data. The results indicated that broilers fed a starter and finisher diet for 30 and 18 d, respectively, would meet the production and processing constraints. Trace plots showed that the production and processing variables were not very sensitive to the grower diet.

Analysis of the nonlinear dynamics of daily broiler growth and feed intake.

Daily BW velocity (BWV) and acceleration (BWA) of individual birds have been demonstrated to be oscillatory. Daily feed intake velocity (FIV) and acceleration (FIA) were hypothesized to be oscillatory and to have a positive relationship with BWV and BWA, respectively. Forty-eight male broiler chicks were individually caged and provided a commercial starter feed ad libitum for 49 d. BW and feed intake (FI) were measured daily. Experiment 1 confirmed that, on a daily basis, BWV, BWA, FIV, and FIA were oscillatory. There was a positive correlation between BW and FI, BWV and FIV, and BWA and FIA. A Kohonen neural network (KNN) clustered BWV and FIV into two and three sequential phases. BWA and FIA analysis did not make definitive clusters. In experiment 2, it was hypothesized that correlation between BWV and FIV would increase with feeding of grower and finisher rations. It was hypothesized that KNN three phase clusters may provide more biologically ideal times of ration change (TORC) for starter, grower, and finisher rations. For 49 d, five treatments, nine birds per treatment, were fed starter, grower, and finisher rations singly or together with dietary changes according to an industry or KNN-determined TORC. Evaluation was made of BW, FI, and carcass characteristics. No significant mean differences were found. Compared to the industry group, the KNN group demonstrated significantly improved uniformity (i.e., smaller SD) of BW (bled out), FI, dressing percentage, and some of the carcass characteristics. Differences between KNN and industry TORC results might have been related to the length of time the birds were fed the starter, grower, and finisher diets.

Multiple-objective (goal) programming model for feed formulation: an example for reducing nutrient variation.

A multiple-objective programming (MOP) model was applied to the feed formulation process with the objectives of minimizing nutrient variance and minimizing ration cost. A MOP model was constructed for a broiler grower ration (3 to 6 wk) and formulated with a Microsoft Excel solver. Twenty-one ingredients with 17 nutrients were included in the formulation. Amino acids were based on digestible values. The following objectives were considered as soft constraints: (1) meeting the nutrient requirements; (2) meeting the ingredient restrictions; and (3) meeting nutrient ratios, including calcium to phosphorus and the relationship of amino acids to lysine (ideal amino acid ratios). Hard constraints considered were (1) a least-cost ration and (2) minimal nutrient variances for protein, methionine, and lysine. It was found that (1) the MOP model was more flexible in providing a compromise solution than a traditional feed formulation with a linear program, (2) the MOP model was able to handle several conflicting objectives simultaneously as compared to the traditional linear programming approach that could handle only one objective, and (3) the MOP model gave the best compromise solution that would satisfy multiple decision makers when trade-offs were made between the ration cost and minimum variances of protein and methionine. The MOP model is an efficient tool to assist the decision-making process through solving a series of linear/nonlinear programs and by interacting with decision-makers.

Minimal number of chicken daily growth velocities for artificial neural network detection of pulmonary hypertension syndrome (PHS).

Previously, evaluation of the first 2 wk of daily growth velocity with an artificial neural network (ANN) provided an effective noninvasive approach for predicting the susceptibility of broilers to pulmonary hypertension syndrome (PHS). This study was conducted to define the minimum number of days of growth data and the type of ANN required for the best prediction of PHS susceptibility. Four experiments were conducted in which broilers were weighed daily at 0800 h. In Experiment 1, Hubbard male broilers were reared to 50 d of age, with 13 developing PHS and 33 remaining normal (N), for a PHS:N ratio of 13:33. In Experiment 2, ANAK broilers were exposed to cool temperatures (16 to 17 C) from 17 to 42 d of age, resulting in a PHS:N ratio of 16:46 for males. In Experiments 3 and 4, Hubbard male and female chicks from a base population and a PHS-resistant line were exposed to cool temperatures from 17 to 42 d (Experiment 3) or 49 d of age (Experiment 4). The PHS:N ratios were 40:68 for males and 6:96 for females in Experiment 3 and 26:91 for males and 10:58 for females in Experiment 4. Four ANN, back propagation (BP3), Ward back propagation (WardBP), probabilistic (PNN), and general regression (GRNN), were evaluated for their ability to predict PHS in the shortest number of days based on daily growth velocities (BWd+1-BWd). A 100% prediction of PHS and N birds was considered the criterion of success. Starting with 14 d of data, each ANN was trained on daily growth velocity, and the number of predictive days was reduced with each run of the ANN. The best ANN was a GRNN, which correctly diagnosed PHS and N male broilers on 4 and 6 d of growth velocity data for Experiments 1 and 2, respectively. The results were poorer with the BP3, WardBP, and PNN. The diagnostic ability of the neural network was not consistent over all four experiments. In Experiment 2, a minimum of 6 d was required for 100% PHS detection for males. In Experiment 3, the best diagnostic value for males was 93% PHS detection and 100% N detection at 15 d. For females, the 100% PHS detection occurred at a minimum of 8 d. In Experiment 4, males had 100% PHS and N detection at a minimum of 11 d. Females had a 100% PHS and N detection at a minimum of 10 d. An attempt to build a single neural network that would detect PHS susceptibility in Hubbard (Experiment 1) and ANAK (Experiment 2) broilers was unsuccessful. The application (validation) of neural networks between experiments also was not successful (data not presented). However, these studies demonstrate that within a breed or line reared under similar selection pressures for ascites, a GRNN based on the first 14 d of growth velocity can detect, with at least 93% accuracy, broilers susceptible to PHS.

Evaluation of broiler growth velocity and acceleration in relation to pulmonary hypertension syndrome.

An evaluation was made of the relationship between individual daily growth patterns and susceptibility of broiler chickens to pulmonary hypertension syndrome (PHS). In the first experiment, 46 male broilers were weighed for each of 50 d, during which time 13 developed PHS. Three temporal phases (0 to 15, 16 to 35, and 36 to 50 d) of broiler growth velocity and acceleration were examined. Correlation dimensions and Lyapunov exponents suggested evidence of chaos in growth velocity and acceleration, but the absence of detectable differences between broilers in the normal and PHS categories led us to reject the hypotheses that growth is more chaotic in normal broilers than in broilers susceptible to PHS. Growth velocity and acceleration values for mean and SD were statistically evaluated as response variables for each growth phase. Mean values for velocity during the third phase were different between broilers in the normal and PHS categories (velocity: 68.8 vs 48.9 g/d, P = 0.03, respectively) and (acceleration: 0.3 vs -1.4 g/d2, P = 0.07, respectively). The third phase SD (reflecting oscillation for velocity and acceleration) was greater for normal than for PHS birds (velocity: 26.1 vs 21.3 g/d, P = 0.13, respectively; acceleration: 39.7 vs 28.2 g/d2, P = 0.03, respectively). The hypothesis was accepted that normal birds have greater oscillations in growth velocity and acceleration than birds susceptible to PHS. A general regression neural network (GRNN) with genetic adaptive calibration was trained to predict PHS based on individual growth phases and their combinations. Data representing the first, first two, and all three phases of growth were determined to have potential for computerized diagnostic weighing. With the GRNN, birds in all three data sets were successfully classified (100%) with or without PHS. A third hypothesis, therefore, was accepted that artificial neural networks could be used to distinguish the difference between normal broilers and those susceptible to PHS. In the second experiment, only one bird was diagnosed with PHS. Velocity and acceleration neural networks from Phase 1 and Phases 1 and 2 in the first experiment were applied to the growth velocity and acceleration data of Experiment 2. The Phase 1 neural networks were the most promising in that they correctly identified 71.6 and 72.4% of the birds as normal for velocity and acceleration data, respectively. In general, data in the second experiment exceeded the neural network range of training for both velocity and acceleration, which reflected increased oscillation during the second phase of growth.

Effects of photoperiod and melatonin on lymphocyte activities in male broiler chickens.

Understanding the role of the pineal gland in regulating the immune response and the role of photoperiod in influencing pineal gland secretions are becoming increasingly important. The purposes of the present experiments were to investigate the effects of different photoperiod regimens on T- and B-lymphocyte activities in broiler chickens. Next, the influence of different photoperiod regimens on the responsiveness of lymphocytes to melatonin in vitro was examined. The effect of melatonin in vitro on lymphocyte activities was also studied, regardless of the photoperiod received. Finally, the effects of photoperiod on the profiles of different splenocyte cell types were investigated. To study the effect of photoperiod on lymphocyte activities, different photoperiod regimens were used. These were: constant lighting, 23 h light:1 h darkness; intermediate lighting, 12 h light:12 h darkness; and intermittent lighting, 1 h light:3 h darkness. Peripheral blood and splenic lymphocyte activities were tested at 3 and 6 wk of age by performing a mitogen cell-proliferation assay with a polyclonal T-cell mitogen, concanavalin A (Con A), and T-dependent B-cell mitogen, pokeweed mitogen (PWM). To study the effect of photoperiod on the responsiveness of lymphocytes to melatonin in vitro or the effect of melatonin in vitro on lymphocyte activities regardless of photoperiod received, lymphocytes from the chickens that were exposed to the different photoperiod regimens were incubated with mitogen and different concentrations of melatonin. To study the effect of photoperiod on profiles of different cell types, the percentages of splenocyte subpopulations from birds exposed to different photo-periods were determined using flow cytometry with CD4+, CD8+, CD3+, and B-cell markers. The results of these studies indicate that splenic T and B lymphocytes from 6-wk-old chickens grown in intermittent lighting had higher activities than those from chickens grown in constant lighting. Peripheral blood and splenic lymphocytes from chickens raised under constant lighting were more responsive to melatonin in vitro than those from chickens raised under intermittent lighting. This difference in response may be due to lower levels of melatonin in birds receiving constant lighting, making them more sensitive to melatonin in vitro. Melatonin in vitro enhanced the mitogenic response of peripheral blood T lymphocytes from 6-wk-old chickens, splenic T lymphocytes from 3-wk-old chickens, and splenic T and possibly B lymphocytes from 6-wk-old chickens. Finally, intermittent lighting increased the percentages of splenic CD4+, CD8+, and CD3+ cells but not B-cell subpopulations at 6 wk of age, presumably because of increased levels of melatonin in birds receiving intermittent lighting. Our results re-emphasize the importance of melatonin in regulating host immune response; this regulation could be accomplished through exposing broiler chicks to intermittent lighting.

Improving neural network prediction of amino acid levels in feed ingredients.

Artificial neural networks (ANN) were trained to predict the amino acid (AA) profile of feed ingredients. The ANN more effectively identified the complex relationship between nutrients and feed ingredients than linear regression (LR). Three types of ANN (NeuroShell 2): three-layer backpropagation (BP3), Ward Backpropagation (WBP), a general regression neural network (GRNN); and LR (SAS Proc GLM) were used to predict the AA level in corn, soybean meal, meat and bone meal, fish meal, and wheat based on proximate analysis. In contrast to a past study, a variety of alternative ANN training parameters were examined to improve ANN performance. Predictive performance was judged on the basis of the maximum R2 value resulting from all defaults tested. Advanced selection of ANN training parameters led to further improvement in performance, especially within the GRNN architecture. In 34 of 35 ANN developed, the maximum R2 value for each individual AA in each feed ingredient was higher for GRNN than for LR, BP3, or WBP prediction methods. For example, the highest R2 value for Met in corn was 0.32 for LR, 0.40 for 3LBP, 0.51 for WBP, and 0.95 for GRNN analysis. Predictive performance was also improved overall as compared to results of a previous study. For example, corn maximum R2 values (GRNN) for Met, TSAA, and Trp were: 0.78, 0.81 and 0.44, previously, and 0.95, 0.96 and 0.88, in the current study. Current soybean meal maximum R values (GRNN) were: Met, 0.92; TSAA, 0.94; and Lys, 0.90. Current meat and bone mean maximum R2 values (GRNN) were: Met, 0.97; TSAA, 0.97; and Lys, 0.97. The ANN computation is a successful alternative to statistical regression analysis for predicting AA levels in feed ingredients.

Probabilistic neural network prediction of ascites in broilers based on minimally invasive physiological factors.

A Probabilistic Neural Network (PNN) was trained to predict ascites in broilers based on minimally invasive inputs (i.e., physiological factors that do not require the death of the bird). A PNN is a supervised, three-layer, artificial neural network that classifies input patterns (e.g., physiological data) into specific output categories (e.g., ascites or no ascites). The PNN inputs were O2 level in the blood, body weight, electrocardiogram (ECG), hematocrit, S wave, and heart rate of individual birds. These data were from three experiments that have been described previously (Roush et al., 1996a,b). The three data sets were pooled into a combined data set for a total of 170 observations. From the pooled data, a training set (117 birds), a calibration set (17 birds), and a verification set (36 birds) were extracted. The PNN was trained on the training data set. To prevent the PNN from overfitting the training data, the neural network was evaluated on its ability to make correct predictions of the calibration data set. At the point at which the neural network made the highest number of correct classifications for the calibration data set, the trained neural network was saved on the computer. When the PNN was applied to the complete data set, the sensitivity or proportion of the birds with ascites that the PNN correctly diagnosed was 0.97 (75/77 birds). The specificity or proportion of birds that the PNN made a correct diagnosis of not having ascites was 0.98 (91/93 birds). When the PNN was applied to the verification data set, which was not subjected to neural network training, the sensitivity was 0.95 (19/20) and the specificity was 0.88 (14/16 birds). Use of models developed with artificial neural networks may enhance the diagnosis of ascites in broilers. The results may be useful in choosing and developing broiler strains that do not have a propensity for ascites.

Artificial neural network prediction of amino acid levels in feed ingredients.

Artificial Neural Networks (ANN), which are biologically inspired tools, serve as an alternative to regression analysis for complex data. Based on CP or proximate analysis (PA) of ingredients, two types of ANN and linear regression (LR) were evaluated for predicting amino acid levels in corn, wheat, soybean meal, meat and bone meal, and fish meal. The two ANN were a three layer Backpropagation network (BP3), and a General Regression Neural Network (GRNN). Methionine, TSAA, Lys, Thr, Tyr, Trp, and Arg were evaluated and R2 values calculated for each prediction method. Artificial neural network training was completed with NeuroShell 2 using Calibration to prevent overtraining. Ninety percent of the data were used as the input for the LR and the two ANN. The remaining 10% (randomly extracted data) were used to calibrate the performance of the ANN. As compared to LR, the R2 values were largest when PA input and GRNN were used. The BP3 did not consistently improve the R2 values for either CP or PA inputs as compared to LR. Each neural net can be incorporated into a computer or spreadsheet program.

Artificial neural network prediction of ascites in broilers.

An artificial neural network was trained to predict the presence or absence of ascites in broiler chickens. The neural network was a three-layer back-propagation neural network with an input layer of 15 neurons (defining 15 physiological variables), a hidden layer of 16 neurons, and an output layer of 2 neurons (the presence or absence of ascites). Male by-products of a breeder pullet line were brooded at 32 and 30 C during Weeks 1 and 2, respectively. The training set for the neural network consisted of data from birds subjected to cool temperatures (18 C) to induce ascites. After training, the predictive ability of the neural network was verified with two new data sets. The second data set was from birds subjected to cool temperatures (18 C). The third data set was from birds subjected to clamping of the pulmonary artery to simulate the physiological processes involved in ascites (the temperature was 24 C). A comparison was made between laboratory diagnostic results and the neural network predicted ascites incidence. The neural network accurately identified the presence or absence of ascites in the first (training) set. Two false positives and one false positive were identified in the second and third verification sets, respectively. The birds identified as false positives were determined to be in the developmental stages of ascites before the occurrence of fluid accumulation. Artificial neural networks were found to effectively identify broilers with and without ascites.

Effect of day to day variation of dietary energy on residual feed intake of laying hens.

An experiment was conducted that examined the effect of day-to-day variability of dietary energy on feed intake behavior. Three levels of average daily dietary ME composition (2,580, 2,814, and 3,009 kcal/kg) and three levels of day-to-day variability of ME (low, medium, and high) were assigned to nine groups of DeKalb-XL laying hens in a 3x3 factorial design. Each of the nine treatments contained 35 individually caged birds. Daily measurements of feed intake, egg production, and egg mass were taken for each bird in this 56-d experiment. Pre- and postexperiment body weights were taken at 23 and 31 wk of age. Increasing levels of day-to-day variance of ME were associated with increased feed consumption for each mean level of ME (P < 0.01) and increased variability of day-to-day feed intake for the 2,580 kcal/kg mean level (P < 0.01). No significant differences (P < 0.01) in egg production egg mass, or body weight were observed.

A nonlinear dynamical (chaos) approach to the analysis of broiler growth.

Mathematical chaos has been observed in a number of biological areas, suggesting that order can be found in systems previously described as random. Nonlinear analyses were conducted to determine whether periodicity or chaos was evident in the growth responses of broiler chickens. Analyses of the absolute growth rate and growth rate acceleration were conducted for four lines of broilers selected at 14 or 42 d for high or low growth rates (Experiment 1) and for a commercial broiler strain (Experiment 2). Resulting Lyapunov exponents (LE) and correlation dimensions (CD) were statistically evaluated. Time series and return map graphics were analyzed. In both experiments, independence of day-to-day growth responses was indicated by low r2 values. In Experiment 1, there were significant differences between lines in growth rate (low, 9.1 +/- .3; high, 12.9 +/- .5 g/d) and the standard deviation of growth rate (low, 5.8 +/- .2; high 7.3 +/- .3 g/d). There were no significant differences for LE or CD values between lines or day of selection. In general, the positive LE, noninteger values of CD, and return map graphics in both experiments suggested the presence of chaotic dynamics. Evaluation of mathematical chaos in broiler growth may give insight into the dynamics and modeling of growth and diseases associated with growth.