Troubleshooting molecular phylogenetic analyses

The number, size, and scope of phylogenetic analyses using molecular data has increased steadily in recent years. This has simultaneously led to a dramatic improvement in our understanding of phylogenetic relationships and a better appreciation for an array of methodological problems that continue to hinder progress in phylogenetic studies of certain data sets and/or particular parts of the tree of life. This review focuses on several persistent problems, including rooting, conflict among data sets, weak support in trees, strong but evidently incorrect support, and the computational issues arising when methods are applied to the large data sets that are becoming increasingly commonplace. We frame each of these issues as a specific problem to be overcome, review the relevant theoretical and empirical literature, and suggest solutions, or at least strategies, for further investigation of the issues involved.