Accelerating the prioritisation of plant species with underexplored medicinal potential: The pm4mp (Phylogenetic Methods for Medicinal Plants) R package

Societal Impact Statement Medicinal plants used in ethnobotanical traditions to treat or prevent diseases have gained renewed interest for their largely untapped potential in drug discovery. In this study, we developed and tested novel methods to prioritise plant species based on their unexplored medicinal potential. By enabling researchers to target the most promising species, these approaches reduce the time and costs of bioprospecting while increasing the likelihood of identifying beneficial compounds. At the same time, they help minimise the environmental impact associated with research activities. Overall, our findings support more sustainable drug discovery practices and highlight the responsible use of biodiversity to advance human health. Summary Plants are a key source of active compounds, with many drugs derived from them. Various methods have been used to explore the medicinal potential of unexploited taxa, but identifying the most active species remains challenging. Molecular phylogenetics holds promise for plant bioprospecting, but issues remain, especially when dealing with large‐scale phylogeny. This paper presents a workflow that integrates new and existing methods to accelerate the identification and prioritisation of potential medicinal plants, focusing on the most promising regions of a phylogeny and assigning a value to each taxon based on its medicinal potential. We introduce pm4mp , an R package that implements the newly developed methods and is available for free on GitHub. pm4mp provides functionalities for identifying stable hot nodes across multiple analysis replicates, extracting hot trees for a disease of interest, prioritising target species by using new approaches and visualising the results graphically. We demonstrate the usefulness of pm4mp by analysing medicinal plant data on 10 diseases from a public database together with a phylogeny of 30,000+ land plants. Our findings show the effectiveness of the newly proposed methods, which represent a substantial enhancement over the existing approaches for plant bioprospecting.