Multivariate statistical analysis techniques such as principal component analysis (PCA) and partial least squares (PLS) analysis have been extensively used in food chemistry for quality control.1, 2 Adulteration of expensive food oils with inexpensive filler oils and mislabeling of tea and wine varieties to suggest they were grown in high-quality regions can be detected cheaply and quickly using multivariate analysis.1, 2 While these methods can be used to classify samples based on geography and taxonomy and identify outliers to prioritise for research projects, very little research has been conducted into the use of multivariate analysis for natural products chemistry.3
Australian sea sponges from the family Dysideidae were selected for this analysis, as they have been found to generally contain three compound structural classes in two chemotypes: one with peptides produced by the microbiome and terpenoids produced by the animal, and the second solely containing microbial oxygenated polyhalogenated diphenyl ethers (O-PHDEs).4 The hypothesis that these metabolites are produced to aid these sessile animals with defence is supported by their frequently-reported potency in antimicrobial, antifeedant, and anticancer assays.4, 5 While some compounds are unique to a particular sponge species, others may be isolated in multiple genera, which adds to difficulties in making taxonomic determinations for these sponges.5
Griffith University’s natural product repository NatureBank contains 30,000 biota samples, including over 200 Dysideidae extracts.6 As a proof-of-concept for this research project, the 1H NMR spectra of 52 Dysideidae extracts from NatureBank were collected for PCA. The resultant model highlighted groups of chemically (and often taxonomically) related extracts and provided guidance into selecting extracts which were more likely to contain interesting compounds. While the model was able to separate sponges with peptides containing thiazole moieties from other samples, there was an insufficient number of O-PHDE containing sponges to allow for chemotype separation.
Keywords: natural products, extracts, sea sponge, Dysideidae, bioactivity, principal component analysis