Multivariate characterization

Bolck, A. and Smilde, A. K., Multivariate characterization of RP-HPLC stationary phases, in Retention and Selectivity in Liquid Chromatography, Smith, R. M., Ed., Elsevier Science, Amsterdam, 1995, chap. 12. 

Fig. 8.6. Representation of 14 multivariately characterized objects in a two-dimensional space of variables where the clusters are connected into four groups (a) and classified into two differently chosen groups (b,c), respectively d shows a nested clustering of B within A... 

Fig. 17.1. Multivariate characterization with VolSurf descriptors. Molecular Interaction Fields (MIF shaded areas) are computed from the 3D-molecular structure. MIFs are transformed in a table of descriptors, and statistical multivariate analysis is performed.

Another use of PCA in multivariate characterization is the formulation of a class model. If there are several classes of subjects in a study, a PC model can be made of each class with surrounding tolerance volumes. New subjects are assigned to a class if it is inside the tolerance volume of this class. This simple but efficient classification scheme is called SIMCA (soft independent modelling of class analogy) and it is described in detail elsewhere [17, 18]. 

In the following discussion an example is given of the multivariate characterization of soil metal status to distinguish pollution of soil layers from geogenic enrichments from the background. Further case studies and a deeper description are given in detail in the literature [KRIEG and EINAX, 1994],... 

P. Goodford, Multivariate Characterization of Molecules For QSAR Analysis,... 

Sandberg, M., Eriksson, L., Jonsson, J., Sjostrom, M., Wold, S. New chemical descriptors relevant for the design of biologically active peptides. A multivariate characterization of 87 amino acids,... 

Chiorboli, C, Piazza, R., Carassiti, V., Passerini, L. and Tosato, M.L. (1993a). Application of Chemometrics to the Screening of Hazardous Substances. Part II. Advances in the Multivariate Characterization and Reactivity Modelling of Haloalkanes. Chemom.IntellLab.Syst., 19, 331-336. 

Eriksson, L., Verhaar, H.J. and Hermens, J.L. (1994b). Multivariate Characterization and Modeling of the Chemical Reactivity of Epoxides. Environ. ToxicolChem., 13,683-691. 

Eriksson, L., Verhaar, H.J.M. and Mermens, J.L.M. (1994) Multivariate characterization and modeling of the chemical reactivity of epoxides. Environ. Toxicol. Chem., 13, 683-691. 

Goodford, P.J. (1996) Multivariate characterization of molecules for QSAR analysis. J. Chemom., 10, 107-117. 

Skagerberg, B., Sjdstrdm, M. and Wold, S. (1987) Multivariate characterization of amino acids by reversed phase high pressure liquid chromatography. Quant. Struct. -Act. Relat., 6, 158-164. 

L Wallbacks, U Edlund, B Norden, and I Berglund. Multivariate Characterization of Pulp Using Solid-State C(13)-NMR, FTIR, and NIR [Near-IR Spectroscopy]. Tappi J. 74 201-206, 1991. 

In the general case, cluster design is applied to a number of candidate molecules, as in D-optimal design, but on a different criterion of representativeness. However, the same clustering approach can also be used for selecting the representative items of a multivariate characterization of discrete systems. 

In the first case, the structural description of over 100 thienyl- and furyl-benzimidazoles and benzoxazoles was multivariately characterized to identify three latent variables. A set of 16 informative molecules was derived thereafter on applying a central composite design criterion in these latent variables to all the available structures. The data were analyzed by a linear PLS model, which permitted the optimization of three structural features out of four. The fourth one, the substituent linked to the homocyclic ring of the bicyclic system was finally optimized by the CARSO procedure in terms of the substituents PPs, predicting two new compounds as possible optimal structures. Indeed, later analysis revealed the accuracy of these predictions. 

In the second case, we took into account over 400 quinolones reported in the literature. Again, a chemometric approach based on multivariate characterization and design in the resulting latent variables permitted us to select a set of 32 molecules with a well-balanced structural variation on which to derive the QSAR models. Linear PLS modeling allowed ranking of the relative importance of individual structural features, and, by CARSO analysis, a new class of compounds was predicted, the lead of which was tested and shown to be as active as expected. This preliminary lead, after a proper modification, is presently being tested for further development. 

An important feature of linear models over non-linear ones is that they usually are easier to interpret. However, the value of a model, expressed in terms of predictivity and interpretability, crucially depends on the type and number of descriptors which are incorporated. 3D-QSAR models based on MlFs allow to localize regions with favorable and less favorable inferacfions. Very often, predictive models are obtained by using multivariate approaches with many descriptors. The choice of descriptors defermines fhe chemical variabilify which can be captured by a QSAR model [135]. During the model building process one has to find an appropriate balance between few, easily inferprefable variables, and a broader multivariate characterization which maybe more difficult to interpret. 

As most often the structural factors determining biological activities are not directly known, multiparameter tables should be used together with a multivariate characterization approach. 

WaUbacks L., Edlund U., Norden B., Berglund I., Multivariate characterization of pulp using soUd-state carbon-13 NMR,... 

Sarra-Bournet, C, Poulin, S., Piyakis, K., Turgeon, S., Laroche, G. (2010) ToF-SIMS multivariate characterization of surface modification of polymers by N2H2 atmospheric pressure dielectric barrier discharge. Surf. Interface Anal., 42,102-109. 

---