Data, methods and assumptions in phylogenetic inference

Abstract

We consider data, methods and assumptions in relation to phylogenetic inference under two main themes, that of assumptions and models relating to methodology (independent of the data), and assumptions and models relating to kinds of data (independent of the methods). Some aspects of methodological assumptions are well known, e.g. those contrasting cladistics and phenetics, but more detailed study of methodological assumptions is needed in relation to analysis of quantitative data and to analysis of distance data. Debate over assumptions relating purely to methodology has perhaps overshadowed considerations of assumptions relating to data. Pattern based evaluation of data includes congruence/consensus measures, iterative weighting schemes and calculation of statistics such as consistency index. These strategies are complementary to recent attempts to build more process based models, for example in the rationale for weighting of transversions over transitions in analysis of sequence data. Recent developments in significance testing bridge the gap between pattern based and process based models. In both of these contexts biological knowledge and interpretation will play an important role.