Partial least-squares regression: a tutorial

A tutorial on the partial least-squares (PLS) regression method is provided. Weak points in some other regression methods are outlined and PLS is developed as a remedy for those weaknesses. An algorithm for a predictive PLS and some practical hints for its use are given.     

Partial least-squares path modelling with latent variables

A partial least-squares treatment of multivariate data related through a complex model allows simultaneous evaluation of the interactions between large numbers of features. Results are given for a model in which water sources flow together; each source is represented by water quality data to allow the influence of the various sources to be evaluated with respect to their importance on the resulting flow downstream.     

Partial least-squares regression and fuzzy clustering — A joint approach

Chemical and physical analyses of malt, the main ingredient of beer, have been used to predict the concentration of certain volatile compounds in the finished beer.The prediction was done by means of the partial least squares regression (PLS2) in SIMCA. The total data set as well as individual malt clusters were submitted to PLS analysis. Best prediction was obtained by separating the total object matrix in classes according to similarity found by fuzzy pattern recognition (FCV). FCV was also used to separate the beer variables in classes and to select the subset of variables to be predicted.A joint approach of fuzzy pattern recognition to identify groups of samples and SIMCA-PLS2 to predict several dependent variables is suggested as a powerful tool in process-analytical chemistry.     

Information content of data and variables and types of weighting in least-squares regression methods

Algorithms are given for evaluating the relative amount of useful information related to a particular parameter which is carried by individual data points and intervals of the variables. The algorithms provide an efficient means of using the information contained in a set of data. Applications to the optimization of weighting in regression methods are described. Several informational and combined informational-statistical types of weighting are studied as a means of improving the accuracy and precision of the parameters obtained by non-linear regression.     

Partial least-squares modelling of dye fastness to light

Several physicochemical properties (¹³C-n.m.r., spectrophotometric, chromatographic, calorimetric) were measured for a series of 38 azo dyes in order to test the existence of relationships between one or more of these data sets and the fastness of the dyes on polyester fabric. The partial least-squares method was applied; it had previously proved useful in establishing structure/property relationships for the same dyes. It is shown that ¹³C-n.m.r. data can be used successfully to predict dye fastness to light. These data are quickly and cheaply obtained with small amounts of compound.     

Kinetic wavelength-pair method as an alternative approach to simultaneous multi-component analysis

The kinetic wavelength-pair method involves adapting a well established principle to the simultaneous kinetic-based determination of two or more compounds with diode-array detection. It relies on measuring the difference in the rate of change of the absorbance with time at two preset wavelength pairs such that the values of these measuring parameters are the result of the contribution of one or two components. The theoretical basis adapted to first-order kinetic-based determinations was established and checked on four systems. The method thus developed was compared with other differential kinetic methods and applied to the resolution of formaldehyde-acrolein mixtures.     

Bioelectrosynthesis as an alternative to photosynthesis

The CO₂ reduction processes have been discussed as a way of designing an ecologically totally closed technology. An electric current and molecular hydrogen are the two related available agents that can be discussed as ecologically pure reductants. The most important products are liquid and gaseous fuels, the products of large-scale organic synthesis, monomers, and amino acids. For CO₂ reduction, the necessary energy consumption and H₂ costs were calculated. For complex organic molecules, amino acids for instance, the energy consumption does not make up the main portion of the costs. The biocatalytic systems of CO₂ reduction based on cryoimmobilized cells are described. Conversion of CO₂ into L-lysine with electrochemical decomposition of water was effected on the laboratory scale. A general unit for diverse technological processes can be a bioelectrosynthetic Index Entries: Bioelectrosynthesis; CO₂ reduction; liquid fuels; amino acids; immobilized cells; economic estimates. modulus, an electrochemical hydrogen generator coupled with a biocatalytic converter of hydrogen and oxygen. The systems for bioelectrosynthesis of motor fuels and essential amino acids have been economically estimated and characterized. The possibilities of combining the solar energy transformation and H₂–CO₂ conversion have been discussed.     

Zwitteration as an alternative to PEGylation

A direct, head-to-head comparison of the efficacy of a zwitterionic versus a poly(ethylene glycol), PEG, coating in preventing protein adsorption to silica and aggregation of silica nanoparticles is presented. The same siloxane coupling chemistry was employed to yield surfaces with similar coverages of both types of ligand. Nanoparticle and planar surfaces were challenged with salt, serum, lysozyme, and serum albumin at 25 and 37 °C. While both types of surface modification are highly effective in preventing protein adsorption and nanoparticle aggregation, the zwitterion provided monolayer-type coverage with minimal thickness, whereas the PEG appeared to yield a more three-dimensional coating. The mechanism for adsorption resistance is thought to be based on preventing ion pairing between protein and surface charges, which releases counterions and water molecules, an entropic driving force enough to overcome a disfavored enthalpy of adsorption.     

Partial least-squares regression for the simultaneous determination of aluminum and beryllium in geochemical samples using xylenol orange

A partial least-squares regression (PLS) was developed for the simultaneous spectrophotometric determination of aluminum and beryllium in geochemical samples by xylenol orange as the chromogenic reagent in water media and in micellar media. The effects of pH, xylenol orange concentration and surfactant concentration on the complexes formation reactions were studied. In the absence of surfactant the complex color development takes place at 2 h after mixing the reagents. By using CTAB as micellar media the complexes formation was improved and the rate of the complexation reactions increased. By using PLS-1 algorithm satisfactory results were obtained. The proposed procedure showed to be useful for prediction of Al and Be values from 0.1 to 1.0 mg L(-1).     

The effects of errors in measuring the independent variable in least-squares regression analysis

It is well known that linear regression analysis gives unbiased estimates of the slope and intercept of a straight line if the dependent variable y is subject to random errors of measurement while the independent variable x is not. It is much less well known that, if x is also subject to random errors of measurement, linear regression analysis yields an underestimate of the slope and a correspondingly biased estimate of the intercept. These errors cannot be removed by weighted regression. Similar errors arise in non-linear regression when the independent variable is afflicted by random errors of measurement.     

On the solution of the least-squares local energy variance minimization equations

The convergence behavior of an iterative procedure proposed by Frost, Kellogg, and Curtis for solution of variance minimization equations is discussed. Use of particular sets of initial parameters in the iterative procedure is shown to yield rapid convergence.     

Partial least-squares modeling of continuous nodes in Bayesian networks

In Bayesian networks it is necessary to compute relationships between continuous nodes. The standard Bayesian network methodology represents this dependency with a linear regression model whose parameters are estimated by a maximum likelihood (ML) calculation. Partial least-squares (PLS) is proposed as an alternative method for computing the model parameters. This new hybrid method is termed PLS-Bayes, as it uses PLS to calculate regression vectors for a Bayesian network. This alternative approach requires storing the raw data matrix rather than sequentially updating sufficient statistics, but results in a regression matrix that predicts with higher accuracy, requires less training data, and performs well in large networks.     

Reinforced molecular recognition as an alternative to rigid receptors

In theory, a perfectly rigid receptor will probably be an unbeatable binder. However, rigidity may not be easy to achieve in practice and it is certainly not Nature's method to realise high affinity. In many proteins binding affinity is increased through non-covalent interactions within the protein. Thus there is a considerable incentive to follow Nature's example and start exploring the use of secondary intra-receptor interactions to aid in the binding process. Secondary interactions within a receptor will reinforce host-guest binding when the same conformational rearrangement (or freezing of motion) is required for guest binding as for the formation of the intra-receptor interactions. Introducing secondary interactions will require rather elaborate synthetic receptors to be produced. With the recent developments in dynamic combinatorial chemistry, access to the desired structures should be facilitated. Whether or not this approach will develop into a practical method remains to be established, but even if it does not, efforts along these lines will lead to a better understanding of the complex interplay between molecular recognition, folding and dynamics.     

Independent component analysis yields chemically interpretable latent variables in multivariate regression

This work shows that independent component analysis (ICA) can be used to obtain statistically independent and, therefore, chemically interpretable latent variables (LVs) in multivariate regression. Two novel algorithms based on ICA are introduced and compared with two classical methods on simulated data: principal component regression and partial least-squares regression. All methods compared yield accurate predictions, but only those based on ICA yield LVs that are chemically interpretable. Practical limitations of ICA-based regression with respect to the underlying assumptions, sample size, and measurement noise are discussed and illustrated by means of simulations.     

Photochemistry of p-chloromercurobenzophenone: Photoreduction as an alternative to demercuration

p-Chloromercurobenzophenone exhibits dual reactivity during photolysis: in solvents that are hydrogen donors it is photoreduced with the formation of a mercury-containing ketyl radical, while in other solvents it undergoes demercuration. Photolysis of p-chloromercurobenzophenone by both paths occurs from the triplet excited state.Translated from Izvestiya Akademii Nauk SSSR, Seriya Khimicheskaya, No. 8, pp. 1738–1742, August, 1990.     

Poly(N-phenylenebenzimidazoles) as an alternative to classical polybenzimidazoles

A new thermally stable polybenzimidazole film has been synthesized by nucleophilic aromatic substitution, transformed into a complex with phosphoric acid and tested as a proton-conductive membrane.     

Shear-flow-based chromatographic separations as an alternative to pressure-driven liquid chromatography

It is only by developing specially designed injection and detection systems that shear-driven chromatography can become a viable alternative to HPLC. In the present paper, a dedicated zero dead-volume injection procedure is presented with which sample volumes can be injected reproducibly in the required picoliter range. In addition, a transversal detection groove system is designed which should allow to perform on-line UV-VIS absorption measurements with path lengths in the millimeter range, with an acceptable theoretical plate loss (only 20% in a 5 cm long channel) and acting as a nearly perfect wave guide.     

Assessment of acetone as an alternative to acetonitrile in peptide analysis by liquid chromatography/mass spectrometry

Acetonitrile as a solvent used in liquid chromatography/mass spectrometry (LC/MS) of peptides and proteins is a relatively toxic solvent (LD₅₀ oral; rat; 2,460 mg/kg) compared to alternatives like methanol (LD₅₀ oral; rat; 5,628 mg/kg) and acetone (LD₅₀ oral; rat; 5,800 mg/kg). Strategies to minimize its consumption in LC are either to reduce the inner diameter of the column or replace acetonitrile with a suitable alternative. Methanol is often recommended to replace acetonitrile in peptide analysis. In this study however, the main focus lies on another alternative solvent for LC/MS of peptides; acetone. A number of model proteins were tryptically digested and the peptide solutions were analyzed on a linear trap quadrupole (LTQ) mass spectrometer. The performances of acetonitrile, methanol and acetone were compared according to the quality of the chromatograms obtained and identification of the peptides using the BioWorks? software developed by Thermo Scientific. In accordance to the elutropic series, acetone was found to significantly reduce the retention times of peptides separated by C18 column material with regard to acetonitrile while methanol led to increased retention times. Acetone was the superior solvent to methanol for most of the tested model proteins reaching similar sequence coverage and numbers of identified peptides as acetonitrile. We therefore propose acetone as an alternative to acetonitrile in LC/MS of peptides. Copyright © 2009 John Wiley & Sons, Ltd.