Weighted-Interaction Nestedness Estimator (WINE): A new estimator to calculate over frequency matrices

We propose a new nestedness estimator that takes into account the weight of the interactions, that is, it runs over frequency matrices. A nestedness measurement is calculated through the average distance from each matrix cell containing a link to the cell with the lowest marginal totals, in the packed matrix, using a weighted Manhattan distance. The significance of this nestedness measure is tested against a null model that constraints matrix fill to observed values and retains the distribution of number of events.This is the first methodological approach that allows for the characterization of weighted nestedness. We have developed a graphical user interface (GUI) running in Matlab to compute all these parameters. The software is also available as a script for R-package and in C++ version.