Heterocovariance Based Metabolomics as a Powerful Tool Accelerating Bioactive Natural Product Identification

The discovery of new active natural products is hampered by laborious purification processes that often end up to the re-isolation of known compounds. We demonstrate here that, spectral data reflecting concentration fluctuations of components can correlate statistically with measurable dose-dependent properties on the basis of a Heterocovariance approach deconvoluting the active components structure. Variance of extract constituents was achieved through statistically meaningful collections of plants from different families, genus, and species. This fluctuation was also obtained through the fractionation of a single plant extract by separation techniques. The NMR and HRMS spectra of the extracts and fractions were recorded, as well as their ability to inhibit tyrosinase or 5-lipoxygenase enzymes. Biological activity was statistically correlated with spectral data deciphering the active compounds through the Heterocovariance approach prior to any purification.