Corno, M. | bioRxiv Science

bioRxiv2020

Supervised and Unsupervised Classification of Cocoa Bean Origin and Processing using Liquid Chromatography-Mass Spectrometry

Liquid Chromatography-Mass Spectrometry (LC-MS) provides an unprecedented wealth of metabolomics information for food products, including insights into compositional changes during food processing. Here, we employed the largest available LC-MS dataset of around 300 cocoa bean samples to assess the capability of two popular multivariate classification methods, principal component analysis (PCA) and linear decomposition analysis (LDA), for studying bean geographic origin and responsible characteristic compounds. The unsupervised method, PCA, only provides a limited separation in bean origin. Expectedly, the supervised method, LDA, provides a better origin clustering. However, it suffers from a strong, nonlinear dependence on the set of compounds used in the analysis. We show that for LDA a compound filtering criterion based on Gaussian intensity distributions dramatically enhances origin clustering of samples, thus increasing its predictive efficiency. In this form, the supervised method of LDA holds the possibility to identify potential markers of a specific origin.

bioinformatics↗

bioRxiv2020

Correlation network analysis based on untargeted LC-MS profiles of cocoa reveals processing stage and origin country

In order to implement quality control measures and create fine flavor products, an important objective in cocoa processing industry is to realize standards for characterization of cocoa raw materials, intermediate and finished products with respect to their processing stages and countries of origin. Towards this end, various works have studied separability or distinguishability of cocoa samples belonging to various processing stages in a typical cocoa processing pipeline or to different origins. Limited amount of success has been possible in this direction in that unfermented and fermented cocoa samples have been shown to group into separate clusters in PCA. However, a clear clustering with respect to the country of origin has remained elusive. In this work we suggest an alternative approach to this problem through the framework of correlation networks. For 140 cocoa samples belonging to eight countries and three progressive stages in a typical cocoa processing pipeline we compute pairwise Spearman and Pearson correlation coefficients based on the LC-MS profiles and derive correlation networks by retaining only correlations higher than a threshold. Progressively increasing this threshold reveals, first, processing stage (or sample type) modules (or network clusters) at low and intermediate values of correlation threshold and then country specific modules at high correlation thresholds. We present both qualitative and quantitative evidence through network visualization and node connectivity statistics. Besides demonstrating separability of the two data properties via this network-based method, our work suggests a new approach for studying classification of cocoa samples with nested attributes of processing stage sample types and country of origin along with possibility of including additional factors, e.g., hybrid variety, etc. in the analysis.

systems biology↗