Validation R2 Research Articles

Validation of fatty acid predictions in milk using mid-infrared spectrometry across cattle breeds

The aim of this study was to investigate the accuracy to predict detailed fatty acid (FA) composition of bovine milk by mid-infrared spectrometry, for a cattle population that partly differed in terms of country, breed and methodology used to measure actual FA composition compared with the calibration data set. Calibration equations for predicting FA composition using mid-infrared spectrometry were developed in the European project RobustMilk and based on 1236 milk samples from multiple cattle breeds from Ireland, Scotland and the Walloon Region of Belgium. The validation data set contained 190 milk samples from cows in the Netherlands across four breeds: Dutch Friesian, Meuse-Rhine-Yssel, Groningen White Headed (GWH) and Jersey (JER). The FA measurements were performed using gas–liquid partition chromatography (GC) as the gold standard. Some FAs and groups of FAs were not considered because of differences in definition, as the capillary column of the GC was not the same as used to develop the calibration equations. Differences in performance of the calibration equations between breeds were mainly found by evaluating the standard error of validation and the average prediction error. In general, for the GWH breed the smallest differences were found between predicted and reference GC values and least variation in prediction errors, whereas for JER the largest differences were found between predicted and reference GC values and most variation in prediction errors. For the individual FAs 4:0, 6:0, 8:0, 10:0, 12:0, 14:0 and 16:0 and the groups’ saturated FAs, short-chain FAs and medium-chain FAs, predictions assessed for all breeds together were highly accurate (validation R2 > 0.80) with limited bias. For the individual FAs cis-14:1, cis-16:1 and 18:0, the calibration equations were moderately accurate (R2 in the range of 0.60 to 0.80) and for the individual FA 17:0 predictions were less accurate (R2 < 0.60) with considerable bias. FA concentrations in the validation data set of our study were generally higher than those in the calibration data. This difference in the range of FA concentrations, mainly due to breed differences in our study, can cause lower accuracy. In conclusion, the RobustMilk calibration equations can be used to predict most FAs in milk from the four breeds in the Netherlands with only a minor loss of accuracy.

Runoff and Sediment Modeling Using SWAT in Gumera Catchment, Ethiopia

This study was undertaken to examine the applicability of the SWAT model in Gumera river basin upstream of Lake Tana, Ethiopia for simulating stream runoff and sediment load. The area of river basin was discretized into 24 sub-catchments using ArcSWAT interface of the model. The semi automated Sequential Uncertainty Fitting (SUFI2) and fully automated Parameter Solution (ParaSol) calibration process built in SWAT calibration and uncertainty program (SWAT-CUP) were used to calibrate the model parameters using time series of flow and sediment load data of 1994 to 2002 and validated with the observed data from years 2003 to 2006. The performance of the model was evaluated using statistical and graphical methods to assess the capability of the model in simulating the runoff and sediment yield for the study area. The coefficient of determination (R2) and NSE values for the daily runoff by using [ParaSol] optimization technique was obtained as 0.72 and 0.71 respectively for the calibration period and 0.79 and 0.78 respectively for the validation period, R2 and NSE values of monthly flow calibration using SUFI2 are 0.83 and 0.78 respectively for validation it was 0.93 and 0.93. For monthly sediment yield by using SUFI2 calibration technique the model evaluation coefficients R2 and NS for calibration was computed as 0.61 and 0.60 respectively, for validation it was 0.84 and 0.83 respectively. The sensitivity analysis on 13 runoff producing parameters was also carried out and discussed.

Comparing near and Mid-Infrared Reflectance Spectroscopy for Determining Properties of Malagasy Soils, Using Global or LOCAL Calibration

Nowadays, near infrared (NIR) and mid-infrared (mid-IR) reflectance spectroscopy are recognised useful approaches for quantifying soil properties, cost and time effectively. The aim of this work was to compare predictions of soil carbon (C) and nitrogen (N) content, C/N ratio, substrate-induced respiration (SIR) and denitrifying enzyme activity (DEA) using NIR and mid-IR spectroscopy over a diverse set of 360 Malagasy topsoils. Partial least square regression was used for fitting NIR and mid-IR spectra to conventional data through procedures of calibration either global (one prediction model for all samples) or LOCAL (one prediction model per sample). Prediction accuracy was assessed according to validation ( r2), standard error of prediction ( SEP) in proportion to the mean and ratio of standard deviation to SEP ( RPD). Using both NIR and mid-IR spectroscopy, global calibration over the whole sample set yielded predictions that were excellent for C and N ( r2 &gt; 0.9, SEP &lt;20%, RPD ⩾ 3), good for C/N, acceptable for SIR, but poor for DEA. LOCAL calibration improved C/N and SIR predictions with both NIR and mid-IR spectroscopy, while DEA prediction became acceptable with NIR spectroscopy only. Additional improvement was achieved when LOCAL calibration was carried out over the fine-textured sub-set, especially for SIR ( r2 &gt; 0.9, SEP &lt; 20%, RPD &gt; 3). In contrast, LOCAL calibration over the coarse-textured sub-set was clearly not useful for improving prediction accuracy. NIR outperformed mid-IR spectroscopy whatever the variable, the calibration procedure and the sample set (except for SIR over the coarse-textured sub-set, where both similar), suggesting its possible superiority for tropical soils.

Near‐infrared spectroscopy for analysis of oil content and fatty acid profile in almond flour

AbstractAlmond kernels show large variability for oil content and fatty acid profile. The objective of this research was to evaluate the potential of near infrared (NIR) reflectance spectroscopy (NIRS) for the analysis of these traits in almond flour. Ground kernels of 181 accessions collected in 2009 were used for developing calibration equations for oil content and concentrations of individual fatty acids. Calibration equations were developed using second derivative transformation and modified partial least squares regression. They were validated with samples from 179 accessions collected in 2010. The accuracy of calibration equations was measured through the coefficient of determination (r2) in external validation and the ratio of the SD in the validation set to the standard error of prediction (RPD). Both r2 and RPD were high for oil content (r2 = 0.99; RPD = 9.24) and concentrations of oleic (r2 = 0.97; RPD = 5.37) and linoleic acids (r2 = 0.98; RPD = 7.35), revealing that calibration equations for these traits are highly accurate. Conversely, the accuracy of the calibration equations for palmitic (r2 = 0.54; RPD = 1.41) and stearic acids (r2 = 0.52; RPD = 1.44) was too low for allowing their application in practice. NIRS discrimination of oil content and concentrations of oleic and linoleic acids was mainly based on the spectral region from 2240 to 2380 nm.Practical applications: NIRS is a high‐throughput analytical technique that allows fast measurement of several traits in a single analysis without using chemical reagents. We evaluated the feasibility of analyzing oil content and concentrations of palmitic, stearic, oleic, and linoleic acids in almond flour using fruits collected during 2 years from a world germplasm collection. The fruits collected in 2009 were used for NIRS calibration, whereas the fruits collected in 2010 were used for validation. NIRS equations were highly accurate for measuring oil content and concentrations of oleic and linoleic acids, which are important traits defining the quality of almond flour for specific uses in the food industry. These results have applications both in the research laboratory and the food industry, where NIRS is becoming a widely used technique for quality control.

Topological Features in Profiling the Antimalarial Activity Landscape of Anilinoquinolines: A Multipronged QSAR Study

The antimalarial activity of a series of 4‐anilinoquinolines was modeled with topological and other functional descriptors using feature selection approaches CP‐MLR and GA. Five models were identified from each approach to explain the activity of the compounds. They jointly shared eighteen descriptors. Among them five descriptors, namely, H‐052, MATS4m, MATS7e, Mor30p, and R7m, were common to both approaches. In PLS analysis the eighteen descriptors have led to a three‐component model (r2 = 0.731, Q2 = 0.688, ). and the common descriptors were among the most influential ones to modulate the activity. Among them, MATS7e indicated the favorability of nonlinear and branched molecular topology for higher activity. MATS4m has also advocated in favor of branching/nonlinearity in the molecule for the activity. The H‐052 argued that R′CH2‐CHX‐CH2R fragments (X is halogen) in the scaffold enhance the activity. In BP‐ANN these descriptors led to very good predictive models (training r2 &gt; 0.81; validation r2 &gt; 0.81; test r2 &gt; 0.75). The study has offered direction to understand the patterns of the antimalarial activity of anilinoquinolines for exploring potential prototype compounds.

Model-Based Integrated Methods for Quantitative Estimation of Soil Salinity from Hyperspectral Remote Sensing Data: A Case Study of Selected South African Soils

Soil salinization is a land degradation process that leads to reduced agricultural yields. This study investigated the method that can best predict electrical conductivity (EC) in dry soils using individual bands, a normalized difference salinity index (NDSI), partial least squares regression (PLSR), and bagging PLSR. Soil spectral reflectance of dried, ground, and sieved soil samples containing varying amounts of EC was measured using an ASD FieldSpec spectrometer in a darkroom. Predictive models were computed using a training dataset. An independent validation dataset was used to validate the models. The results showed that good predictions could be made based on bagging PLSR using first derivative reflectance (validation R2 = 0.85), PLSR using untransformed reflectance (validation R2 = 0.70), NDSI (validation R2 = 0.65), and the untransformed individual band at 2 257 nm (validation R2 = 0.60) predictive models. These suggested the potential of mapping soil salinity using airborne and/or satellite hyperspectral data during dry seasons.

Prediction of water quality parameter in Jajrood River basin: Application of multi layer perceptron (MLP) perceptron and radial basis function networks of artificial neural networks (ANNs)

River water quality is a significant concern in many countries, considering agricultural and drinking consumptions. Therefore, prediction of salinity index, as the main water quality condition is a necessary tool for water resources planning and management. This paper describes the application of artificial neural networks (ANNs) models for computing the total dissolved solids (TDS) level in Jajrood River (Iran). Two ANN networks, multi-layer perceptron (MLP) and radial basis function (RBF), were identified, validated and tested for the computation of TDS concentrations. Both networks employed five input water quality variables measured in river water over a period of 40 years. The performance of the ANN models was checked through the coefficient of determination (R2) and root mean square error (RMSE). Jajrood River is one of the most important rivers which is located adjacent to Tehran city and supplies drinking water for people who live in this mega-city and recreational uses. Tehran is the most populous city and largest industrial pole in Iran, which caused the river, to be exposed to various pollutants. Matlab 2007 was selected for modeling goals in this research. Results show that MLP and RBF modeling as two methods of ANN are able to simulate water quality variables of Jajrood River with more than 90% accuracy. After modeling in MLP and RBF formatting and comparing simulation results (output) show that, the RBF result (R2 of validation is 0.9362) are more closely to reality than the MLP (R2 of validation is 0.8968). In other words, because of large number of input data, the RBF modeling performance has a better prediction than MLP modeling. Key words: Water quality variables, artificial neural network (ANN), multi-layer perceptron (MLP), radial basis function (RBF).

2-D QSAR model development for α-amino suberic acid derivatives as a novel anticancer agent

Structure activity relationship has been reported for 34 compounds derived from α-amino suberic acid. Compounds in this class typically inhibit human histone deacetylase by hyperacetylation of histone protein in both normal and cancer cells induce expression of p21 and differentiate surviving cancer cells to a non-proliferating phenotype. In the present study, quantitative structure activity relationship study was performed on a series of novel α-amino suberic acid derivatives as inhibitors of histone deacetylase. Multiple linear regression analysis was performed to derive quantitative structure activity relationship models which were further evaluated internally as well as externally for the prediction of activity. The best quantitative structure activity relationship model for multiple linear relationship analysis was selected with a correlation coefficient (r 2) of 0.855, cross validation r 2 (CV) 0.809 and the standard error 0.287. Further partial least square and neural network analysis were also performed for partial least square analysis correlation coefficient (r 2) 0.876, CV r 2 0.756, standard error 0.287 for neural networking analysis R 2 of training set 0.861, R 2 of test set 0.661. All of them generated comparable results which prove that the model formed is sound and has good predictability.

Use of generalized additive models and cokriging of spatial residuals to improve land-use regression estimates of nitrogen oxides in Southern California

Land-use regression (LUR) models have been developed to estimate spatial distributions of traffic-related pollutants. Several studies have examined spatial autocorrelation among residuals in LUR models, but few utilized spatial residual information in model prediction, or examined the impact of modeling methods, monitoring site selection, or traffic data quality on LUR performance. This study aims to improve spatial models for traffic-related pollutants using generalized additive models (GAM) combined with cokriging of spatial residuals. Specifically, we developed spatial models for nitrogen dioxide (NO2) and nitrogen oxides (NOx) concentrations in Southern California separately for two seasons (summer and winter) based on over 240 sampling locations. Pollutant concentrations were disaggregated into three components: local means, spatial residuals, and normal random residuals. Local means were modeled by GAM. Spatial residuals were cokriged with global residuals at nearby sampling locations that were spatially auto-correlated. We compared this two-stage approach with four commonly-used spatial models: universal kriging, multiple linear LUR and GAM with and without a spatial smoothing term. Leave-one-out cross validation was conducted for model validation and comparison purposes. The results show that our GAM plus cokriging models predicted summer and winter NO2 and NOx concentration surfaces well, with cross validation R2 values ranging from 0.88 to 0.92. While local covariates accounted for partial variance of the measured NO2 and NOx concentrations, spatial autocorrelation accounted for about 20% of the variance. Our spatial GAM model improved R2 considerably compared to the other four approaches. Conclusively, our two-stage model captured summer and winter differences in NO2 and NOx spatial distributions in Southern California well. When sampling location selection cannot be optimized for the intended model and fewer covariates are available as predictors for the model, the two-stage model is more robust compared to multiple linear regression models.

3D-QSAR CoMFA and CoMSIA studies for design of potent human steroid 5α-reductase inhibitors

3D-QSAR studies were performed on a set of sixty-one 6-azasteroidal human steroid 5α-reductase inhibitors. The models developed using comparative molecular field analysis (CoMFA) and comparative molecular similarity indices analysis (CoMSIA) methodologies employing atom and centroid based alignment were reliable and significant having good predictive r 2 value. CoMSIA model developed by a combination of steric, electrostatic, hydrophobic and hydrogen bond donor showed good relative and predictive properties. The predictive power of the models was assessed using an external test set of 10 compounds. The best model had shown cross validation r 2 (0.702) with 7 optimum components, non-cross-validation r 2 (0.956), F value (86.158), predicted r 2 (0.704), standard error of estimate (0.165). Further the CoMFA/CoMSIA contour plots show a preference toward the C-17 substituents and indicate that bulky group at C-17 is an important requirement for exhibiting 5α-reductase inhibition. These developed models could be used to design and optimise the more potent 6-aza steroidal inhibitors of 5α-reductase enzyme. The 3D-QSAR studies involving comparative molecular field analysis and comparative molecular similarity indices analysis on a series of 6-azasteroids as 5α-reductase inhibitors is reported.

Pharmacophore Identification and QSAR Studies on Substituted Benzoxazinone as Anti-Platelet Agents: kNN-MFA Approach

The three-dimensional quantitative structure–activity relationship (3D-QSAR) and pharmacophore identification studies on 28 substituted benzoxazinone derivatives as antiplatelet agents have been carried out. Multiple linear regression (MLR) method was applied for QSAR model development considering training and test set approaches with various feature selection methods. Stepwise (SW), simulated annealing (SA) and genetic algorithm (GA) were applied to derive QSAR models which were further validated for statistical significance and predictive ability by internal and external validation. The results of pharmacophore identification studies showed that hydrogen bond accepters, aromatic and hydrophobic, are the important features for antiplatelet activity. The selected best 3D kNN-MFA model A has a training set of 23 molecules and test set of 5 molecules with validation (q2) and cross validation (pred_r2) values 0.9739 and 0.8217, respectively. Additionally, the selected best 3D QSAR (MLR) model B has a training set of 23 molecules and test set of 5 molecules with validation (r2) and cross validation (pred_r2) values of 0.9435 and 0.7663, respectively, and four descriptors at the grid points S_123, E_407, E_311 and H_605. The information rendered by 3D-QSAR models may lead to a better understanding and designing of novel potent antiplatelet molecules.

Predicting bovine milk protein composition based on Fourier transform infrared spectra

Phenotypic information on individual protein composition of cows is important for many aspects of dairy processing with cheese production as the center of gravity. However, measuring individual protein composition is expensive and time consuming. In this study, we investigated whether protein composition can be predicted based on inexpensive and routinely measured milk Fourier transform infrared (FTIR) spectra. Based on 900 calibration and 900 validation samples that had both capillary zone electrophoresis (CZE)-determined protein composition and FTIR spectra available, low to moderate validation R2 were reached (from 0.18 for αS1-casein to 0.56 for β-lactoglobulin). The potential usefulness of this model on the phenotypic level was investigated by means of achieved selection differentials for 25% of the best animals. For α-lactalbumin (R2=0.20), the selection differential amounted to 0.18g/100g and for casein index (R2=0.50) to 1.24g/100g. We concluded that predictions of protein composition were not accurate enough to enable selection of individual animals. However, for specific purposes when, for example, groups of animals that meet a certain threshold are to be selected, the presented model could be useful in practice on the phenotypic level. The potential usefulness of this model on the genetic level was investigated by means of genetic correlations between CZE-determined and FTIR-predicted protein composition traits. The genetic correlations ranged from 0.62 (β-casein) to 0.97 (whey). Thus, predictions of protein composition, when used as input to estimate breeding values, provide an excellent means for genetic improvement of protein composition. In addition, estimated repeatabilities based on 3 repeated observations of predicted protein composition showed that a considerable amount of prediction error can be removed using repeated observations.

Quantificação de matéria orgânica do solo através de modelos matemáticos utilizando colorimetria no sistema Munsell de cores

O presente estudo teve como objetivo desenvolver modelos matemáticos para a quantificação do teor de matéria orgânica, a partir da cor do solo, obtida por aparelho colorímetro no sistema Munsell de cores. Para esse fim, 912 amostras de solo foram coletadas na região de Porto Grande (Amapá) e enviadas para análises química, granulométrica e determinação da cor em amostras secas e úmidas. Os componentes valor e croma da cor do solo no sistema Munsell, obtidos por colorímetro, foram utilizados para quantificar através de regressão múltipla passo a passo (stepwise) o teor de matéria orgânica do solo. O modelo de predição com base em todas as amostras apresentou R² de 0,66 para amostras úmidas e 0,56 para amostras secas, ao serem validados utilizando amostras independentes. Foi possível ainda melhorar os modelos quando as amostras foram separadas por classe de solo ou textura, e os modelos gerados com base em cores de amostras úmidas foram sistematicamente superiores àqueles utilizando amostras secas. Em relação às classes de solo, os melhores resultados foram obtidos para Argissolos e Latossolos, ambos gerando um R² de validação independente de 0,73 (amostra úmida). Para textura, os melhores resultados foram obtidos para solos de textura muito argilosa, com R² de validação de 0,81 (amostra úmida). Os modelos de predição de matéria orgânica em função da cor do solo possuem simplicidade e potencial para serem utilizados no laboratório e no campo, especialmente para Argissolos e Latossolos de textura argilosa, de maneira automática e sem necessidade de uso de produtos ou reagentes.

EFFECT OF THE NUMBER OF MEASUREMENT SITES ON LAND USE REGRESSION (LUR) MODELS OF AIR POLLUTANTS

Background and Aims: Land use regression (LUR) models used in epidemiologic studies on air pollution usually rely on air pollution measurements at a few locations across a city. The number of locations commonly varies from 20 to 100. Our aim is to explore the effect of the number of locations (N) on LUR model performance. Methods: The REGICOR-AIR study measured NO2 in 159 locations in the two adjacent cities of Girona and Salt (Spain). A total of 102 potential predictors like traffic, road type and land cover entered a supervised forward selection procedure to derive the final model. Training datasets of 20, 50, 80, 110 and 140 points were obtained by randomly sampling the original dataset. The remaining points were used as validation datasets. The process was repeated 200 times for each N. Adjusted R(Ra) and leave-one-out cross-validation R(Rcv) were calculated in the training datasets, and validation R2 was obtained from the validation datasets. The median value over the 200 iterations was reported. Results: With 20 points, the median R2a and R2cv were 79% and 71% but the validation R2 was negative, indicating that the regression equation resulted in a worse prediction than assigning the overall mean to each point of the validation sample. Both R2a and R2cv decreased with increasing N to reach R2a=54% and R2cv=52% for N=140. Conversely, the validation R2 increased with increasing N, from negative values (N=20) to 44% (N=140). The most frequent predictive variable entered in only 32% of the models for N=20, while for N=140 four variables appeared in more than 90% of the final models. Conclusions: Increasing the number of measurement points increases both the predictability and the stability of the models. Leave-one-out cross-validation R2 overestimates the predictive ability of the model. This bias is more important at small sample sizes.

Impacts of socio-demographic characteristics on male migrants: Logistic regression approach

The study of migration is the key importance not only in Demography but also in Population Studies. Migration is the primary stage of development. Migration is related to development, urbanization, industrialization as well as technological advancement. The aim of this study is to observe the effects of socio-economic and demographic factors on male migrants. The sample data of 700 respondents was collected by multi-stage sampling technique at Meherpur Sadar thana at Meherpur district, Bangladesh. In this paper, logit regression model has been used to determine the influential factors of migration. The study shows that the most of the male migrants are migrated due to services. The analysis indicated that age, place of birth, occupation, monthly income and land property have significant effects on causes of migration among the selected variables in the case of considering dependent variable as causes of migration. Whereas on the other hand, place of birth, monthly income, type of family and land property have significant effects on the type of migration. Moreover, using the cross validation predictive power (CVPP) criterion and R2 , the shrinkage coefficient (.) is constructed. The shrinkage coefficient determines the adequacy of the model.

Spring flood reconstruction from continuous and discrete tree ring series

This study proposes a method to reconstruct past spring flood discharge from continuous and discrete tree ring chronologies, since both have their respective strengths and weaknesses in northern environments. Ring width or density series provide uninterrupted records that are indirectly linked to regional discharge through a concomitant effect of climate on tree growth and streamflow. Conversely, discrete event chronologies constitute conspicuous records of past high water levels since they are constructed from trees that are directly damaged by the flood. However, the uncertainty of discrete series increases toward the past, and their relationships with spring discharge are often nonlinear. To take advantage of these two sources of information, we introduce a new transfer model technique on the basis of generalized additive model (GAM) theory. The incorporation of discrete predictors and the evaluation of the robustness of the nonlinear relationships are assessed using a jackknife procedure. We exemplify our approach in a reconstruction of May water supplies to the Caniapiscau hydroelectric reservoir in northern Quebec, Canada. We used earlywood density measurements as continuous variables and ice‐scar dates around Lake Montausier in the James Bay area as a discrete variable. Strong calibration (0.57 &lt; 0.61 &lt; 0.75) and validation (0.27 &lt; 0.44 &lt; 0.58) R2 statistics were obtained, thus highlighting the usefulness of the model. Our reconstruction suggests that, since ∼1965, spring floods have become more intense and variable in comparison with the last 150 years. We argue that a similar procedure can be used in each case where discrete and continuous tree ring proxies are used together to reconstruct past spring floods.

On variable relations between vegetation patterns and canopy reflectance

Statistical relations between the species composition of plant assemblages and canopy reflectance are frequently employed in remote sensing for mapping vegetation at local scales. Reflectance is influenced by species composition but also affected by dynamics such as seasonal vegetation development or plant stress. Due to this variability in time and space, doubts are frequently raised with respect to the transferability of statistical relations in remote sensing of plant assemblages. Hence, this study addresses the stability of statistical relations between species composition and reflectance despite of spatiotemporally changing vegetation conditions. We established permanent plots at three temperate sites (nutrient-poor grassland, wet heath, and floodplain meadow). We measured canopy reflectance at multiple dates over the vegetation period using a field spectrometer with hyperspectral resolution. Simultaneously, plant species composition and other vegetation and surface parameters that may exert influence on reflectance were recorded. Species composition was statistically related to the corresponding reflectance data using ordination (Isometric Feature Mapping) and cross-validated regression models (Partial Least Squares Regression). Time series of model fits as well as regression coefficients were used to estimate the temporal stability of the models. Model fits were further compared to changes in vegetation conditions. Model residuals were tested for co-variable influences. Finally, we tested the transferability of the statistical relations in time. Results showed that species composition could be modeled with rather high accuracies (R² in validation up to 0.78 and for only three measurements lower than 0.5), with the highest fits near the vegetation optimum (i.e., the date with maximum cover of photosynthetically active vegetation). The transferability in time varied with the vegetation type. Uncertainties in the models were strongly related to variable canopy height and to the occurrence of litter. Since such spatial heterogeneities may be a result of non-stationary processes, we conclude that statistical methods taking into account such effects may further improve the accuracy of vegetation mapping.

Quantitative Measurements of Loss on Ignition in Iron Ore Using Laser-Induced Breakdown Spectroscopy and Partial Least Squares Regression Analysis

Laser-induced breakdown spectroscopy (LIBS) and partial least squares regression (PLSR) have been applied to perform quantitative measurements of a multiple-species parameter known as loss on ignition (LOI), in a combined set of run-of-mine (ROM) iron ore samples originating from five different iron ore deposits. Global calibration models based on 65 samples and their duplicates from all the deposits with LOI ranging from 0.5 to 10 wt% are shown to be successful for prediction of LOI content in pressed pellets as well as bulk ore samples. A global independent dataset comprising a further 60 samples was used to validate the model resulting in the best validation R(2) of 0.87 and root mean square error of prediction (RMSEP) of 1.1 wt% for bulk samples. A validation R(2) of 0.90 and an RMSEP of 1.0 wt% were demonstrated for pressed pellets. Data preprocessing is shown to improve the quality of the analysis. Spectra normalization options, automatic outlier removal and automatic continuum background correction, which were used to improve the performance of the PLSR method, are discussed in detail.

The effect of the number of observations used for Fourier transform infrared model calibration for bovine milk fat composition on the estimated genetic parameters of the predicted data

Fourier transform infrared spectroscopy is a suitable method to determine bovine milk fat composition. However, the determination of fat composition by gas chromatography, required for calibration of the infrared prediction model, is expensive and labor intensive. It has recently been shown that the number of calibration samples is strongly related to the model's validation r2 (i.e., accuracy of prediction). However, the effect of the number of calibration samples used, and therefore validation r2, on the estimated genetic parameters of data predicted using the model needs to be established. To this end, 235 calibration data subsets of different sizes were sampled: n=100, n=250, n=500, and n=1,000 calibration samples. Subsequently, these data subsets were used to calibrate fat composition prediction models for 2 specific fatty acids: C16:0 and C18u (where u=unsaturated). Next, genetic parameters were estimated on predicted fat composition data for these fatty acids. Strong relationships between the number of calibration samples and validation r2, as well as strong genetic correlations were found. However, the use of n=100 calibration samples resulted in a broad range of validation r2 values and genetic correlations. Subsequent increases of the number of calibration samples resulted in narrowing patterns for validation r2 as well as genetic correlations. The use of n=1,000 calibration samples resulted in estimated genetic correlations varying within a range of 0.10 around the average, which seems acceptable. Genetic analyses for the human health-related fatty acids C14:0, C16:0, and C18u, and the ratio of saturated fatty acids to unsaturated fatty acids showed that replacing observations on fat composition determined by gas chromatography by predictions based on infrared spectra reduced the potential genetic gain to 98, 86, 96, and 99% for the 4 fatty acid traits, respectively, in dairy breeding schemes where progeny testing is practiced. We conclude that a relatively large number of calibration samples is required to be able to obtain genetic correlations that lie within a limited range. Considering that the routine recording of infrared spectra is relatively cheap and straightforward, we concluded that this methodology provides an excellent means for the dairy industry to genetically alter milk fat composition.

Greater Sage‐Grouse Nesting Habitat: The Importance of Managing at Multiple Scales

Abstract: Considering habitat selection at multiple scales is essential to fully understand habitat requirements and management needs for wildlife species of concern. We used a hierarchical information‐theoretic approach and variance decomposition techniques to analyze habitat selection using local‐scale habitat variables measured in the field and landscape‐scale variables derived with a Geographic Information System (GIS) for nesting greater sage‐grouse (Centrocercus urophasianus) in the Powder River Basin (PRB), Montana and Wyoming, USA, 2003–2007. We investigated relationships between habitat features that can and cannot be mapped in a GIS to provide insights into interpretation of landscape‐scale—only GIS models. We produced models of habitat selection at both local and landscape scales and across scales, yet multiscale models had overwhelming statistical and biological support. Variance decomposition showed that local‐scale measures explained the most pure variation (50%) in sage‐grouse nesting‐habitat selection. Landscape‐scale features explained 20% of pure variation and shared 30% with local‐scale features. Both local‐ and landscape‐scale habitat features are important in sage‐grouse nesting‐habitat selection because each scale explained both pure and shared variation. Our landscape‐scale model was accurate in predicting priority landscapes where sage‐grouse nests would occur and is, therefore, useful in providing landscape context for management decisions. It accurately predicted locations of independent sage‐grouse nests (validation R2 = 0.99) and showed good discriminatory ability with &gt;90% of nests located within only 40% of the study area. Our landscape‐scale model also accurately predicted independent lek locations. We estimated twice the amount of predicted nesting habitat within 3 km of leks compared to random locations in the PRB. Likewise we estimated 1.8 times more predicted nesting habitat within 10 km of leks compared to random locations. These results support predictions of the hotspot theory of lek placement. Local‐scale habitat variables that cannot currently be mapped in a GIS strongly influence sage‐grouse nest‐site selection, but only within priority nesting habitats defined at the landscape scale. Our results indicate that habitat treatments for nesting sage‐grouse applied in areas with an unsuitable landscape context are unlikely to achieve desired conservation results.