A comparative study concerning the chromatographic behaviour and lipophilicity of certain natural toxins

Abstract

The characterization and classification of some natural toxins (mycotoxins and alkaloids) is established by high-performance thin-layer chromatography (HPTLC) using chemically bonded stationary phases of RP-18, RP-18W, RP-8, RP-2, RP-NH(2) , RP-CN and RP-Diol. The used mobile phases are mixtures of methanol-water in different proportions of volume, chosen to create a suitable migration during the chromatographic process. The lipophilicity is described through different chromatographic descriptors, such as R(M0), mean of R(M) (mR(M)), and scores of R(M) values corresponding to the first principal component (PC1/R(M)). The investigated compounds are characterized also by a series of computed lipophilicity descriptors (logP and logD values). The chromatographic behaviour of toxins is evidenced by graphical profiles of the lipophilicity descriptors. In addition, scores and loadings resulting from covariance matrix of retention data enable new information about similarity and differences observed between the investigated compounds. The experimental lipophilicity indices estimated from retention data are directly correlated with the computed values, via computer software and Internet module, at a high significant statistical level.

Multivariate assessment of lipophilicity scales—computational and reversed phase thin-layer chromatographic indices

“…For this type of problems, correlation matrices are most often used. Pearson's correlation coefficient has been already extensively used for the assessment of novel lipophilicity indices [24,25,44]. In order to extend the capabilities of SRD to this type of problems, we have implemented a MS Excel VBA macro for the generation of "SRD heatmaps", where we use all of the variables in turn as the reference to produce a matrix of SRD values.…”

“…Instead of the current misuse of Pearson's correlation matrices [24,25,44] we would like to encourage the implementation of: (a) adequate arrangement of variables and (b) coloring schemes which enables patterns among variables to be easily perceivable. The choice of the best variable/lipophilicity measure following the aforementioned matrix arrangement is straightforward, i.e.…”

“…Aside from very specific applications of chromatographic approaches that tend to mimic biosystems such as micellar liquid chromatography (MLC) [10][11][12][13][14][15], immobilized artificial membrane chromatography (IAM) [16,17], immobilized proteins etc. [18], the mainstream methods in the determination of lipophilicity are still based on typical reversed-phase chromatography including a variety of chemically bonded stationary phases [19][20][21][22], where octyl-, octadecyl-, and cyanopropyl-modified silica beds are the most frequently used in combination with a polar mobile phase (usually binary mixtures of miscible organic solvents and water) [23][24][25].…”

“…the first principal component scores (PC1/R M ) derived from principal component analysis (PCA) of multivariate retention data [30,31], and arithmetic means of R M values, more frequently denoted as mR M [23][24][25].…”

“…Many procedures use similarities among computationally estimated values and experimentally derived lipophilicity indices as a criterion to select the best one. Such similarities are most often obtained from hierarchical clustering (HCA) [43,44], principal component analysis [21,23,25,45], or simple correlations based on parametric statistics such as Pearson's correlation coefficient [24,25,44]. The last one is applicable only if the data is normally distributed, which is often not the case.…”