Planta Med 2012; 78 - PG15
DOI: 10.1055/s-0032-1320658

Mutual information analysis of salvia exudates

S Bertolini 1, A Bisio 2, G Rauch 3, E Giacomelli 2, G Mele 2, M Giacomini 1
  • 1Department of Communication Computer and System Science, University of Genoa, Via All'Opera Pia 13, Genoa I-16145, Italy
  • 2Department of Pharmacy, University of Genoa, Via Brigata Salerno, Genoa I-16147, Italy
  • 3Institute of Biophysics, CNR, National Research Council, Via De Marini 6, Genoa I-16149, Italy

As Salvia exudates in a screening work have shown to possess inhibitory activity on the germination of Papaver rhoeas L. and Avena sativa L., a database for fast and efficient data collection among the various collaborating groups to be used in the following research steps to define the allelopathic potential of these species has been developed. Data about the effect of 13 species of Salvia (13 species and a control species) on Avena and Papaver has been extracted from the database and some statistical analysis have been done. First, an Artificial Neural Network (ANN) clustering algorithm has been developed in order to group data in at most five clusters. 100 trials have been done to cluster data, in order to have unbiased results. Then, the clustering results have been ranked by the following innovative index, to give a ranking of the effectiveness of every Salvia species:

Nt is the number of clusters in the trial t; cte is the cluster in which the element e is present in the trial t; NCit is the number of neurons of the cluster in which the species i is present in the trial t; NTit is the total number of neurons where the species i is present in the trial t. Finally, a Mutual Information analysis has been done in order to decide what number of clusters could be the one that leads to the best classification. The best one has found to be 3.