Prediction of Bioactive Metabolites from American Aconitum Using Network Integrating Cellular Morphological Profiling and Mass Spectrometry Data.

Asian and American Aconitum species are phylogenetically close, but only certain Asian species have been well-studied for their medicinal properties. This study aims to discover bioactive compounds in two American Aconitum species based on a systematic networking strategy integrating both mass spectrometry data and biological profiles from a high-throughput phenotypic screening assay, Cell Painting. The chemical profiles of four different plant parts of two American Aconitum species (A. columbianum and A. uncinatum) were obtained by ion mobility mass spectrometry and compared with two Asian (A. carmichaelii and A. fischeri), and one European species (A. napellus). Biological screening, image analysis, and feature extraction were performed on Aconitum extracts using the Cell Painting assay. The results provided 2,090 unique morphological features per extract, which were further reduced to 429. In conjunction with 4,400 chemicals from a library with known mechanisms of action, 198 unique hierarchical clusters were established. An overall activity heuristic called CP score was calculated for each sample. After integrating the CP score and spectrometric data, a network filtered for higher CP scores was constructed and the compounds with high activity were tentatively identified. The network contained mostly American Aconitum species, suggesting that these understudied plants produce useful bioactive compounds.