Format: ORAL

Giovanni Zecca1,2, Elisa Toini1,2, Massimo Labra1,2, Fabrizio Grassi1,2

1 Department of Biotechnology and Biosciences, University of Milano-Bicocca, Milan 20126, Italy 2 NBFC, National Biodiversity Future Center, Palermo 90133, Italy

Recent researches have proposed that phylogenetic methods can be effective tools for exploiting the untapped potential of medicinal plants. Under the assumption that medicinal properties tend to be phylogenetically clustered, phylogenies can be used to narrow down the relevant species for bioprospecting, increase the discovery rate and reduce the time and costs involved. Nonetheless, some restrictions exist. By integrating ethnobotanical knowledge with phylogenetic methods, it has often been necessary to group plants into macro-categories of therapeutic effects, limiting the potential of the approach. In many studies hot nodes corresponding to lineages overrepresented in medicinal taxa were identified. However, as larger phylogenies and databases become available, the possibility of huge lineages being identified by hot nodes is also likely to increase, once again posing a prioritization problem. To tackle these issues, we propose a new analysis pipeline, mixing existing and new developed methods. Briefly, our workflow is as follows: 1) using selected databases that integrate traditional medicine data with growing molecular information, species linked to specific diseases or specific biological effects are obtained; 2) different methods to estimate the phylogenetic clustering of taxa are applied using a reference phylogeny; 3); where a phylogenetic signal has emerged, “hot nodes" are repeatedly estimated to assess their robustness; 4) in case of large lineages subtended by stable “hot nodes”, a (relative) probability of being linked to the disease or biological effect of interest is calculated for each taxa; 5) results are summarised and graphically displayed to combine different information. To test our pipeline data related to 10 diseases and 10 biological effects were obtained from two different public databases and were analysed together with a published phylogeny including ~ 30,000 species. Our results show the effectiveness of the approach suggesting that it will be useful in the bioprospecting of new potential medicinal plants.