Reduction In Residual Variance Research Articles

Mixed Models in Nonlinear Regression for Description of the Growth of Nelore Cattle.

Body weight records were used to characterize the growth curve of Nelore cattle. Body weight was regressed as a function of age, for both sexes, by using nonlinear models through the functions of Brody, Gompertz, Logistic, Richards, Meloun 1, Von Bertalanffy, and Von Bertalanffy. The quality of the model arrangements was evaluated by employing Akaike and Bayesian Schwarz information criteria. The Brody function provided the best adaptations by the evaluators and, considering the asymptotic weight and the maturation rate as random, a reduction in residual variance of 79% for males and 83% for females was obtained in relation to the models under fixed contexts. In males, the absolute and relative growth rates ranged from 0.921 to 0.261 kg/day and 2.39 to 0.08%, respectively. For the same rates, under another approach, females ranged from 0.922 to 0.198 kg/day and 2.55 to 0.06%, respectively. Males showed greater growth acceleration at the beginning of the growth trajectory, being equal to females at 397 days of age and from that age onward they presented lower estimates. The nonlinear regression model approach under the mixed-models context allows reduction of residual variance, increasing model accuracy.

Additive and non-additive genetic effects of humoral immune traits in Japanese quail

In breeding programs, using appropriate models to estimate the variance components with high accuracy is essential. Considering the non-additive genetic effects along with additive effects in evaluation analysis models can be effective in increasing the accuracy of estimates. This study was conducted to fit the best model for estimation of variance components of humoral immune responses (antibody titers against sheep red blood cells [SRBC] and Newcastle disease virus [NDV]), especially non-additive genetic effects (dominance and epistasis) in Japanese quail. The collected records of humoral immune responses were used to estimate genetic parameters using the Bayesian method via Gibbs sampling by fitting 24 different models. The best model for each trait was selected based on Deviance Information Criteria (DIC). Moreover, the residual and additive genetic variance changes were evaluated comparing the models. The results showed that dominance, epistasis, and maternal genetic effects were influenced on humeral immune responses traits except for immunoglobulin Y (IgY). The direct heritability for total antibody titer (AbT), the titer of immunoglobulin Y (IgY), the titer of immunoglobulin M (IgM) against SRBC, antibody titer against NDV (AbNDV), and immune system performance index (IgF) were 0.065, 0.208, 0.117, 0.100, and 0.132, respectively. The contribution of dominance and epistasis as a proportion of phenotypic variance for all traits ranged from 0.070 to 0.221, and 0.087 to 0.251, respectively. A reduction in residual variance were observed in all traits, moreover accuracy of estimated breeding values increased in some traits with adding non-additive genetic effects in the models. Furthermore, the additive genetic variance in the best models decreased for AbT, IgF, and AbNDV traits compared with the simple model (without non-additive genetic effects). The results suggested including the non-additive genetic effects in the genetic models for evaluation of humoral immune responses of quail to obtain precise genetic parameters.

Heritability estimates of meat sensory characteristics are a function of the number of panellists and their inter-correlations

The objective of the present study was to quantify, using simulated data, the impact on estimated heritability of varying the number of panellists and their inter-correlations using meat sensory tenderness in cattle as an example. Estimated parameters from actual sensory-based tenderness scores from 9 individual panellists on 1252 beef cattle were used to parameterise the simulation. A single “tenderness score” for each of 10 panellists was simulated for 15,000 cattle. Heritability estimates were calculated for each of the 10 panellists individually as well as the mean score per animal for all n combinations of panellists. Heritability estimates improved with increasing number of panellists in line with expectations from a deterministic equation. The increase in heritability was due to a reduction in the residual variance, albeit the rate of reduction in residual variance declined with each additional panellist included in the calculated mean tenderness score. Results highlight the importance of reporting the number of panellist scores per animal as well as their inter-correlations in sensory-based studies.

Bending properties and strength grading of Norway spruce: variation within and between stands

Current strength grading of Norway spruce (Picea abies (L.) Karst.) structural timber is only able to describe parts of the great variability in density and bending properties. This study assesses whether information about the origin of the timber can be used to predict its strength and stiffness, alone or in combination with machine strength grading. Three hundred and seventy-three boards from 45 trees sampled from three stands in eastern Norway were studied. Substantial parts of the variability of density, modulus of elasticity (MOE), and bending strength or modulus of rupture (MOR) of the boards studied were explained by origin (differences between sites, relative tree size (diameter at breast height), and longitudinal position in stem). Origin also gave a reduction in residual variance in addition to what was obtained by machine grading based on resonance frequencies. For MOR, the improvement was larger than what was obtained by adding density, whereas for MOE, the density was more important than information about origin.

Space Suit Bioenergetics: Cost of Transport During Walking and Running

Space suits significantly affect the metabolic cost of human movement during extravehicular activity, a key component of future human missions to the Moon and Mars. We hypothesized that space suit legs act as springs during running, thereby lowering cost of transport relative to space-suited walking. We transformed data from suited and unsuited energetics studies into a common format and developed a regression equation for specific resistance, S, (energy per weight per distance) based on the Froude number (a nondimensional parameter associated with gait transitions), surface slope, gravity, and space suit pressure. Acceptance criteria for regression factors included significance and a reduction in residual variance. We divided suited data into two groups, running and walking or slow running, and performed a group means hypothesis test and categorical regression of metabolic rate per unit weight and S, measures of time- and distance-associated metabolic efficiency, respectively. The regression on S achieved an adjusted R2 of 0.83; all factors were significant (P < 0.0005). No additional evaluated factors met the acceptance criteria. The categorical regression, but not the hypothesis test, suggested that the running group had reduced efficiency per unit time; both tests suggested that the running group had increased efficiency per unit distance. The regression model largely explained variations in S across studies. Several findings suggest that gas-pressure space suit legs function as springs during running, including the finding of higher efficiency per unit distance during running, despite the presumed increased work against space suit joint torques at higher velocities.

Marked Effect of β-Lactoglobulin Polymorphism on the Ratio of Casein to Total Protein in Milk

The relationship between genetic variants for milk protein and the composition of milk was analyzed on 4475 repeated milk samples from individual cows; 371 dairy cows of the Swedish Red and White breed and 204 cows of the Swedish Holstein breed were used. The registrations included percentages of casein, protein, fat, and lactose in combination with milk yield and SCC. The genotype of individual cows for αs1-CN, β-CN, κ-CN, and β-LG was determined by alkaline and acidic PAGE. A mixed animal model was used for the analysis; β-LG and aggregate casein genotypes were included simultaneously as separate fixed effects in the statistical model. The results suggest a positive additive effect of the β-LG B allele on casein content and on the ratio of casein to total protein. For the latter trait, the β-LG genotype accounted for a relatively large part of the phenotypic variance, corresponding to a reduction in residual variance of 11% when included in the model. The corresponding value for casein content was 0.5%. The lack of unfavorable associations between milk protein variants and the traits included in this study makes the β-LG gene an obvious candidate when the breeding objective is improved conversion of milk protein into cheese.

Optimizing plot size and shape for field experiments on terraces

Two methods, namely combined plot analysis and integration of variograms, have been applied to investigate and compare plot-to-plot yields of intercropped millet (mean yield of about 1.9 t ha−1) and maize (mean yield 4.25 t ha−1) on terraces in Nepal. Combining plots diminishing the residual variance of millet from 0.52 (t/ha)2 of the original 1 m×1 m units to 0.08 (t/ha)2 for 4 m×4 m plots, the largest for which reliable estimates could be derived. The residual variance of maize declined from 4.91 (t/ha)2 of the original units to 0.61 (t/ha)2 for 4 m×4 m plots. Orientation was immaterial. The variogram for millet was isotropic and bounded with a range of approximately 4 m. Within-plot variance calculated from the variogram increased, and experimental error decreased, as the size of plots increased to this value, beyond which there was little gain. The variogram for maize was unbounded, and so gave no ‘natural' upper limit for plot size, that is, over the range studied, up to 6 m×6 m. It was also anisotropic with larger gradient in the direction from front to back of the terrace than along the contour. The reduction in residual variance was greatest for narrow plots elongated in this direction. Formulae are provided from which to calculate from the variogram approximate residual variances for experiments of given area and size, shape, and orientation of plots. Its application gave similar results to those from the combined plot analysis. Overall, the results from the two methods were consistent.

Forecasting the Dutch heavy truck market: A multivariate approach

In this paper it is investigated whether multivariate time series models can improve the forecasting performance of Holt-Winters models and univariate ‘naive’ ARIMA-models, which are easier to construct, but do not take into account lead-lag relationships. This research deals with the Dutch truck market performance in the light of overall economic developments. A 5-variate model is built, containing two truck sales series and three economic indicators. It is found that a substantial reduction in residual variance can be found by using a multivariate model. In the case of one truck series (rigids) this leads to uniformly better forecasts. For the other output series (artics), no improvement in forecasting accuracy was found. Holt-Winters models do not seem to improve any of the sophisticated models.

Three-dimensional P-wave velocity image under the Carpathian arc

An inversion of P-wave travel time residuals from selected earthquakes in the distance range 30°–98° to two seismic station networks was used to model P-wave velocity anomalies down to 250 km depth. In the first inversion experiment a region between 43.5°–47.5°N and 21°–29°E was modelled, using 35 seismic stations, while in the second one a region between 44°–47°N and 25°–29°E was modelled, using 19 seismic stations. The 4-layer block model of the first inversion offers 19% reduction in residual variance, while the 5-layer block model of the second one offers 26% reduction, the rest being explained by noise and smaller scale heterogeneities. The obtained velocity anomalies correlate remarkably well with the gravity anomalies and with the tectonic model for the Vrancea region of Fuchs et al. (1979).

FATTY ACIDS AND STEROIDS AFFECTING FLAVOR AND AROMA OF MEAT FROM RAM, CRYPTORCHID, AND WETHER LAMBS

Organoleptic tests of meat and fat samples from 33 ram, 37 cryptorchid, and 31 wether lambs showed no significant differences in flavor or aroma among the three sex types. There was, however, a small variation in flavor and aroma among the animals. Analyses of variance of flavor and aroma scores revealed that there was a larger Variation within animals (between judges) than that between animals and the repeatability estimates for the two traits were 0.10 and 0.03. However, a significant (P &lt; 0.01) multiple regression of flavor on 15 independent variables (six fatty acids and nine steroids) was determined and its multiple correlation coefficient was 0.56. The reduction in residual variance due to stearic acid was significant (P &lt; 0.05) and linoleic acid was the best predictor of flavor as a single independent variable (P &lt; 0.10). Simple correlation coefficient between flavor and aroma was estimated to be negative and nonsignificant, −0.12.