Principal component regression analysis applied in structure-activity relationships. 1. selective dihydrofolic acid reductase inhibitors

The principal component regression analysis is used when the predicting variables correlate significantly against the regressor variables and, when a collinearity and multicollinearity exist among regressor variables. The example adapted from enzymology shows that no loss of information occurs with respect to the chemical constants in comparison with the reduced model of multivariate regression analysis.     

Correlation between qualitatively distributed predicting variables and chemical terms in acridine derivatives using principal component analysis

When biological variables are not continuously distributed, the multiple and multivariate regression analysis cannot be used to correlate these variables against chemical regressors. As the employment of discriminant analysis requires the homogeneity of dispersion matrices and, that n_hp where n_h= degree of freedom of hypothesis, p =number of chemical terms, the reliability and validity of this method is highly questionable here. An alternative method is based on the principal component analysis where multicategory variables of drug responses can be classified into measures of inactive, slightly active, sufficiently active, and highly active drugs, for instance. The rules for classification are based on biological sources that can be expressed by chemical terms, too. An example adapted from antitumor action of acridine derivatives shows the working technique.     

Q-mode regression analysis applied to predict the antidotal action of bispyridinium oximes in intoxications with organophosphorus poisons

There are some serious pitfalls if the usual multiple and multivariate regression analysis are applied to correlate the biological activity of compounds against parameters describing the chemical molecule structure if the observations consist of an inhomogeneous data set. By contrast to the more known methods of cluster analysis, it must be required that a partition into homogeneous groups is possible and, in addition, that the members inside any group are normally distributed. The principle of the Q-mode regression analysis dealing with that problem is exemplified on the antidotal action of bispyridinium oximes in intoxication with organophosphorus poisons (soman, tabun, sarin, VX).     

Response Surface Analysis for Continous Regressors Applied to Neurotoxic Organophosphorus Pesticides

Usually, the purpose of the response surface optimization is to be able to locate the optimum operating levels of the regressors. In quantitative structure-activity relationship (QSAR) studies, the predicting variable reflects any biological property, and the regressors are structural and physico-chemical terms. By contrast to the Box-Wilson approach, the regressors are continously distributed. The working technique is demonstrated on an example adapted from organophosphorus pesticide research. It was found that the maximum neurotoxicity depends on lipophilic and steric substituent properties.     

Multicollinearity in spatial genetics: separating the wheat from the chaff using commonality analyses

Direct gradient analyses in spatial genetics provide unique opportunities to describe the inherent complexity of genetic variation in wildlife species and are the object of many methodological developments. However, multicollinearity among explanatory variables is a systemic issue in multivariate regression analyses and is likely to cause serious difficulties in properly interpreting results of direct gradient analyses, with the risk of erroneous conclusions, misdirected research and inefficient or counterproductive conservation measures. Using simulated data sets along with linear and logistic regressions on distance matrices, we illustrate how commonality analysis (CA), a detailed variance‐partitioning procedure that was recently introduced in the field of ecology, can be used to deal with nonindependence among spatial predictors. By decomposing model fit indices into unique and common (or shared) variance components, CA allows identifying the location and magnitude of multicollinearity, revealing spurious correlations and thus thoroughly improving the interpretation of multivariate regressions. Despite a few inherent limitations, especially in the case of resistance model optimization, this review highlights the great potential of CA to account for complex multicollinearity patterns in spatial genetics and identifies future applications and lines of research. We strongly urge spatial geneticists to systematically investigate commonalities when performing direct gradient analyses.     

Adjusting for confounding due to population admixture when estimating the effect of candidate genes on quantitative traits

When analyzing the relationship between allelic variability and traits, a potential source of confounding is population admixture. An approach to adjusting for potential confounding due to population admixture when estimating the influence of allelic variability at a candidate gene is presented. The approach involves augmenting linear regression models with additional regressors. Family genotype data are used to define the regressors, and inclusion of the regressors ensures that, even in the presence of population admixture, the estimates of the regression coefficients that parameterize the influence of allelic variability on the trait are unbiased. The approach is illustrated through an analysis of the influence of apolipoprotein E genotype on plasma low density lipoprotein cholesterol concentrations.     

QSAR-based prediction of anti-HCV activity of thiourea derivatives

Quantitative structure–activity relationship (QSAR) studies were performed on a series of thioureas to explore the physico-chemical parameters responsible for their activity against the hepatitis C virus (HCV)-infected AVa5 cell. The physico-chemical parameters were calculated using WIN CAChe 6.1. Multiple linear regression analysis, after the variables selection by factor analysis, was performed to derive QSAR models which were further evaluated for their statistical significance and predictive power by internal and external validation. The developed QSAR model had the correlation coefficient (R) = 0.928 and cross-validated squared correlation coefficient (Q ²) = 0.751. The selected significant QSAR model indicates that hydrophobicity, dielectric energy, valence connectivity index (order 1), conformational minimum energy and highest occupied molecular orbital of the whole molecule play an important role in the anti-HCV activity of thioureas.     

Raman based chemometric model development for glycation and glycosylation real time monitoring in a manufacturing scale CHO cell bioreactor process

The Quality by Design (QbD) approach to the production of therapeutic monoclonal antibodies (mAbs) emphasizes an understanding of the production process ensuring product quality is maintained throughout. Current methods for measuring critical quality attributes (CQAs) such as glycation and glycosylation are time and resource intensive, often, only tested offline once per batch process. Process analytical technology (PAT) tools such as Raman spectroscopy combined with chemometric modeling can provide real time measurements process variables and are aligned with the QbD approach. This study utilizes these tools to build partial least squares (PLS) regression models to provide real time monitoring of glycation and glycosylation profiles. In total, seven cell line specific chemometric PLS models; % mono-glycated, % non-glycated, % G0F-GlcNac, % G0, % G0F, % G1F, and % G2F were considered. PLS models were initially developed using small scale data to verify the capability of Raman to measure these CQAs effectively. Accurate PLS model predictions were observed at small scale (5 L). At manufacturing scale (2000 L) some glycosylation models showed higher error, indicating that scale may be a key consideration in glycosylation profile PLS model development. Model robustness was then considered by supplementing models with a single batch of manufacturing scale data. This data addition had a significant impact on the predictive capability of each model, with an improvement of 77.5% in the case of the G2F. The finalized models show the capability of Raman as a PAT tool to deliver real time monitoring of glycation and glycosylation profiles at manufacturing scale.     

Estimating and controlling for spatial structure in the study of ecological communities

Aim Variation partitioning based on canonical analysis is the most commonly used analysis to investigate community patterns according to environmental and spatial predictors. Ecologists use this method in order to understand the pure contribution of the environment independent of space, and vice versa, as well as to control for inflated type I error in assessing the environmental component under spatial autocorrelation. Our goal is to use numerical simulations to compare how different spatial predictors and model selection procedures perform in assessing the importance of the spatial component and in controlling for type I error while testing environmental predictors. Innovation We determine for the first time how the ability of commonly used (polynomial regressors) and novel methods based on eigenvector maps compare in the realm of spatial variation partitioning. We introduce a novel forward selection procedure to select spatial regressors for community analysis. Finally, we point out a number of issues that have not been previously considered about the joint explained variation between environment and space, which should be taken into account when reporting and testing the unique contributions of environment and space in patterning ecological communities. Main conclusions In tests of species‐environment relationships, spatial autocorrelation is known to inflate the level of type I error and make the tests of significance invalid. First, one must determine if the spatial component is significant using all spatial predictors (Moran's eigenvector maps). If it is, consider a model selection for the set of spatial predictors (an individual‐species forward selection procedure is to be preferred) and use the environmental and selected spatial predictors in a partial regression or partial canonical analysis scheme. This is an effective way of controlling for type I error in such tests. Polynomial regressors do not provide tests with a correct level of type I error.     

Prediction of macrofungal occurrence in Swedish beech forests from soil and litter variable models

The spatial distribution of 49 macrofungal species in Swedish beech forests was related to the statistical variation in 31 edaphic variables. In order to reduce the multicollinearity problem, the variables were transformed into eight principal components, PCs, which are used in two-group discriminant analysis (on absence/presence patterns) and multiple regression analysis (on number of fruit-bodies). The results suggested that base saturation and organic matter content are of outstanding importance. However, significant relationships were also found with other variables, i.e. Cd or Zn in soil and litter, soil nitrogen mineralization rate, and Na or S in litter. One interesting interpretation of the results is that fungi do not only respond to the main variables of a gradient (soil pH, organic mater, base saturation) but also to other variables. Attempts were made to interpret the PCs to characterize fungal occurrence from the models they formed.     

Are clinical parameters sufficient to model gait patterns in patients with cerebral palsy using a multilinear approach?

The aim of this study was to evaluate whether clinical parameters are sufficient using, a multilinear regression model, to reproduce the sagittal plane joint angles (hip, knee, and ankle) in cerebral palsy gait. A total of 154 patients were included. The two legs were considered (308 observations). Thirty-six clinical parameters were used as regressors (range of motion, muscle strength, and spasticity of the lower). From the clinical gait analysis, the joint angles of the sagittal plane were selected. Results showed that clinical parameter does not provide sufficient information to recover joint angles and/or that the multilinear regression model is not an appropriate solution.