GC-MS chemoprofiling, chemotaxonomic differentiation and PASS online prediction of biological activities of essential oils from five Kenyan Hypericum species

Abstract Hypericum species are potential sources of essential oils. Whereas there exists reported work on volatile constituents of Hypericum essential oils (Eos) from other parts of the world, the species of Hypericums in Kenya remain largely unexplored. The current study assessed the chemical variability of Eos from five Hypericum species collected in Kenya using multivariate chemometric approaches. Eos were extracted by hydrodistillation and analyzed using gas chromatography–mass spectrometry (GC–MS). Further, prediction of activity spectra for substances was done. After data cleaning, including merging duplicate compounds, removing absent compounds, and autoscaling, a dataset of 113 compounds was subjected to principal component analysis (PCA), Pearson correlation analysis, hierarchical cluster analysis, phytochemical class profiling, diversity indices, Venn analysis, and discriminating-compound assessment. The volatile profiles showed clear interspecific variation. Sesquiterpenes predominated in Hypericum kiboense , Hypericum roeperianum , Hypericum peplidifolium , and Hypericum scioanum , whereas Hypericum revolutum contained a stronger monoterpene component dominated by α-pinene (43.04%) and caryophyllene (19.40%). Characteristic major constituents included valencene in H. kiboense (28.77%), 6-hydroxy-1-phenylhexane-1,3-dione in H. roeperianum (11.73%), caryophyllene in H. peplidifolium (15.06%), and caryophyllene oxide in H. scioanum (10.68%). The first two principal components explained 68.2% of total variation and clearly separated the species. Correlation analysis revealed generally weak similarities among species, except for the strong relationship between H. peplidifolium and H. scioanum (r = 0.81). Cluster analysis identified three major chemotypic groups. Diversity metrics further highlighted substantial differences in richness, evenness, and overall diversity among species. The five Hypericum species possess distinct but partially overlapping volatile oil fingerprints and are not chemically interchangeable. The results provide a valuable foundation for chemotaxonomy, quality control, and future bioactivity-guided investigations. The predicted biological activity profile showed that antieczematic activity had the highest number of occurrences across the investigated Hypericum essential oils, followed by antineoplastic and apoptosis agonist activities. The ranking table further showed that some less frequent activities, such as spasmolytic urinary activity and mucomembranous protection, had high weighted mean Pa scores, suggesting strong but more narrowly distributed predicted bioactivity.