Interspecific and Environmental Influence on the Foliar Metabolomes of Species Through Recursive OPLSDA Modeling.

Understanding interspecific and environmental influences on secondary metabolite profiles can be critical in plant metabolomics. This study used a hierarchical orthogonal projections to latent structure discriminant analysis (OPLS-DA) to classify the foliar metabolomes of four naturally growing species in Thailand, , , , and . Using a recursive binary classification, interspecific and environmental influences were determined in multiple class separations, while identifying key metabolites driving these distinctions. Gas chromatography-mass spectrometry (GC-MS) annotated 409 metabolites, and through a progressive class differentiation using hierarchical OPLS-DA, exhibited a metabolome distinct from the other three species. However, the metabolomes of and had a lot of overlap, while displayed regional metabolic variation, emphasizing the role of environmental factors in shaping its chemical composition. Key metabolites, such as mitragynine, isorhynchophylline, squalene, and vanillic acid, among others, were identified as major discriminators across the hierarchical splits. Unlike conventional OPLS-DA, which struggles with multiclass datasets, the recursive approach identified class structures that were biologically relevant, without the need for manual pairwise modeling. The results aligned with prior morphological and genetic studies, validating the method's robustness in capturing interspecific and environmental differences, which can be used in high-dimensional multiclass plant metabolomics.