Multivariate color scale image analysis - Thin layer chromatography for comprehensive evaluation of complex samples fingerprint.

A chemometric evaluation of the information provided by different color scale fingerprints in thin layer chromatographic analysis of complex samples is proposed for the correct classification of a set of medicinal plant extracts. The fingerprints of the samples were acquired on HPTLC Silica gel 60 F254 and HPTLC Silica gel 60 plates using multiple levels of visualization under UV light. Images processing on red (R), green (G), blue (B) and respectively grey (K) color scale selection was used in order to evaluate the complete chromatographic profile of the extracts. Combination of Principal Component Analysis (PCA) and Factor Analysis (FA) method was applied in order to reveal the individual contribution of each color scales in the analysis of chromatographic fingerprints. The suggested technique provides an applicable strategy to screen for efficacy-associated color scale for grouping/classification of the extracts exploiting the information provided by HPTLC fingerprints. The principal component analysis and linear discriminant analysis (PCA-LDA) method was applied for the evaluation of numerical data provided by color scale fingerprints digitization and for samples classification. A correct classification of the analyzed extracts according to the plants phylum was revealed by color scale fingerprints analysis. The proposed methodology could be considered as a promising tool with future applications in plant material investigations even from the taxonomic perspective classification.

The impact of the order of derivative spectra on the performance of pattern recognition methods. Classification of medicinal plants according to the phylum.

Data pre-processing is an important strategy in chemometrics and related fields because in many cases the transformation of data has a great effect on the performance of the method (model). However, a careful examination of the literature clearly points out that only very few systematic studies are dedicated to the effect of the derivative spectra on the performance of the pattern recognition methods. This comprehensive study compares the impact of the order of derivative spectra and other data pre-processing procedures (normalization and standardization) on the performance of cluster analysis, principal component analysis and discriminant analysis applied for characterization and classification of medicinal plants according to their phylum using UV spectra. The efficiency of the pre-processing methods was estimated by comparing the accuracy of classification and prediction measured by internal cross-validation. Derivatization method (1st order) resulted in the best classification (100%) of medicinal plants according to their phylum (Pteridophyte, Magnoliophyte and Spermatophyte) as compared to other pre-processing methods (normalized spectra-71.4%, standardized spectra-76.2% and original spectra-78.6%).

A comprehensive classification of edible oils according to their radical scavenging spectral profile evaluated by advanced chemometrics.

A comprehensive study concerning the characterization and classification of 30 cold-pressed edible oils according to their UV-Vis spectra and radical scavenging profiles using the 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay is presented. Considering the principal component analysis (PCA) and fuzzy-principal component analysis (FPCA) loadings profiles, the characteristic spectral regions with a significant influence in oil samples classification were identified and associated with characteristic factors in each group. Much more, the oils with high antiradical capacity were revealed. The scores corresponding to the first principal component and the canonical scores corresponding to the first discriminant function derived from radical scavenging spectral profiles allowed a relevant classification of oils in well-defined groups associated with their high, medium and low radical scavenging capacity. The FPCA-LDA method applied on DPPH radical scavenging spectral profiles of edible oils appeared to be the most efficient method with a correct classification rate of 96.7%.

Classification of Romanian medicinal plant extracts according to the therapeutic effects using thin layer chromatography and robust chemometrics.

Thin layer chromatography in combination with image analysis and advanced chemometric methods were successfully used to classify the medicinal herbs according to their therapeutic effects and usage. The investigations were conducted using two types of plates (HPTLC Silica gel 60 and HPTLC Silica gel 60 F254) which were evaluated in UV light at 254 and 365 nm. The holistic evaluation of the numerical data corresponding different image processing channels (blue, grey, red, green) was performed by employing appropriate multivariate methods: hierarchical cluster analysis (HCA), principal component analysis (PCA), fuzzy principal component analysis (FPCA) and linear discriminant analysis (LDA) applied to the first relevant principal components. The results obtained by applying LDA method indicate a highly accurate separation of the medicinal herbs within the four groups, in good agreement with therapeutic effects and usage. According to this classification, the best image processing channels were identified for each of the investigated HPTLC plates: blue channel for HPTLC Silica gel 60 F254 (with 92.9% percent of discrimination in case of PCA and FPCA) and respectively red channel for HPTLC Silica gel 60 (with 93.9% percent of discrimination in case of FPCA). The 2D and 3D score scatterplots illustrate also the accurate and reliable discrimination between the four distinct groups.