Landmarks: optimizing geographical distances as morphometric information for phylogeographic data.

M. DOLORES CASAGRANDA; J. SALVADOR ARIAS

Honolulu, Hawai, EEUU

Congreso; Hennig XXIX; 2010

Willi Hennig Society

Landmarks: optimizing geographical distances as morphometric information for phylogeographic data. M. Dolores Casagranda & J. Salvador Arias CONICET, INSUE, Facultad de Ciencias e Instituto Miguel Lillo, Universidad Nacional de Tucumán, Miguel Lillo 205, 4000, San Miguel de Tucumán, Tucumán, Argentina Most methods for assignation of ancestral distribution of species population are based on predefined areas, and left out accurate geographical information provided by the data. It is possible to directly optimize geographical information as continuous characters on a phylogenetic tree by using recently described techniques of phylogenetic morphometry (Catalano et al., Cladistics, in press). Assuming that spatial position of terminals (i.e."landmarks") are heritable traits ("homologous characters"), this method is perfectly suitable to haplotype geographical data. Inference of ancestral distribution under parsimony assumes that new generations move as minimum as possible from the geographical area occupied by its immediate ancestors, so geographic position of nodes is optimized to minimize the sum of distances across the tree. Given that the values optimized at nodes are geographical coordinates (latitude‐longitude), the reconstruction can be easily visualized on a map. Here, we show an application of the described alternative, and compare and discussed with previous methods Unlike maximum likelihood methods based on complex movement models, parsimony offers an optimality criterion to map distributional characters onto a single phylogenetic tree, where the geographic distances are not correlated with branch length, so the method would be powerful in cases such as phylogeography of epidemic outbreaks among others.