More characters or more taxa for a robust phylogeny--case study from the coffee family (Rubiaceae).

Using different data sets mainly from the plant family Rubiaceae, but in parts also from the Apocynaceae, Asteraceae, Lardizabalaceae, Saxifragaceae, and Solanaceae, we have investigated the effect of number of characters, number of taxa, and kind of data on bootstrap values within phylogenetic trees. The percentage of supported nodes within a tree is positively correlated with the number of characters, and negatively correlated with the number of taxa. The morphological analyses are based on few characters and weakly supported trees are expected. The percentage of supported nodes is also dependent on the kind of data analyzed. In analyses of Rubiaceae based on the same number of characters, RFLP data give trees with higher percentage of supported nodes than rbcL and morphological data. We also discuss the support values for particular nodes at the familial and subfamilial levels. Two new data sets of ndhF and rbcL sequences of Rubiaceae are analyzed and together with earlier studies of the family we can conclude that the monophyly of the Rubiaceae is supported and within the family there are three well supported, but not easily characterized, large subfamilies, Rubioideae, Cinchonoideae s.s. and Ixoroideae s.l. There are also a few genera (Luculia and Coptosapelta) unclassified to subfamily.