Comparison of partial least squares and random forests for evaluating relationship between phenolics and bioactivities of Neptunia oleracea

• Main Authors: Soo, Yee Lee, Mediani, Ahmed, Maulidiani, M., Khatib, Alfi, Ismail, Intan Safinar, Zawawi, Norhasnida, Abas, Faridah
• Format: Article
• Language: English; English; English
• Published: John Wiley and Sons Ltd 2018
• Subjects: RS403 Materia Medica-Pharmaceutical Chemistry
• Online Access: http://irep.iium.edu.my/59136/1/59136_Comparison%20of%20partial%20least%20squares%20and%20random.pdf; http://irep.iium.edu.my/59136/2/59136_Comparison%20of%20partial%20least%20squares%20and%20random_SCOPUS.pdf; http://irep.iium.edu.my/59136/13/59136_Comparison%20of%20partial%20least%20squares%20and%20random_WoS.pdf; http://irep.iium.edu.my/59136/; https://onlinelibrary.wiley.com/doi/epdf/10.1002/jsfa.8462

BACKGROUND: Neptunia oleracea is a plant consumed as vegetable and used as folk remedy for several diseases. Herein, two regression models (partial least square, PLS and random forest, RF) in metabolomics approach were compared and applied for the evaluation of relationship between phenolics and bioactivities of N. oleracea. In addition, the effects of different extraction conditions on the phenolic constituents were also assessed by pattern recognition analysis. RESULTS: Comparison of the PLS and RF showed that RF exhibited poorer generalization and hence poorer predictive performance. Both the regression coefficient of PLS and the variable importance of RF revealed that quercetin and kaempferol derivatives, caffeic acid and vitexin-2- O-rhamnoside were significant towards the tested bioactivities. Furthermore, principal component analysis (PCA) and partial least square-discriminant analysis (PLS-DA) results showed that sonication and absolute ethanol are the preferable extraction method and ethanol ratio, respectively, to produce N. oleracea extracts with high phenolic levels and therefore high DPPH-scavenging and α-glucosidase inhibitory activities. CONCLUSION: Both the PLS and RF are useful regression models in metabolomics study. This work provides insight into the performances of different multivariate data analysis (MVDA) tools and the effects of different extraction conditions on the extraction of desired phenolics from plant.