My Research

Male Reproductive System
Hymenoptera. Vespina. Vespida. Bees, ants, sawflies, horntails and other wasps. Male reproductive system. Male genitalia of Hymenoptera. Symphyta. Phylogenetic systematics of Hymenoptera. Phylogenetic systematics of Symphyta. Xyeloidea, Tenthredinoidea, Pamphilioidea, Cephoidea, Siricoidea, Xiphydrioidea, Orussoidea. Xyelidae, Blasticotomidae, Tenthredinidae, Diprionidae, Cimbicidae, Argidae, Pergidae, Pamphiliidae, Megalodontesidae, Megalodontidae, Cephidae, Siricidae, Anaxyelidae, Xiphydriidae, Orussidae.


Developmental sequences

Some time ago, I had an idea that solves the long-standing problem of how to include developmental sequences in comparative or phylogenetic analyses. Ward Wheeler refined the idea and incorporated the method in his computer program POY. We also discovered that the currently used method, event-pairing (Smith 1996, Velhagen 1997), is logically flawed -- not only for phylogenetic analysis, but also for comparative analysis. Hence, our method is currently the only valid method for phylogenetic analysis of developmental sequences. Why event-pairing is faulty and how our new method works is explained in Schulmeister & Wheeler (2004).

Robust-choice sensitivity analysis

In phylogenetic analysis of DNA sequences, costs have to be assigned to the six types of substitution and insertion-deletion events (indels). The values of the costs can have a significant influence on the outcome of a cladistic analysis. If only a single phylogenetic analysis is performed with one set of parameter values, it remains completely unknown how much the result is dependent on these arbitrarily chosen values. Sensitivity analysis (Wheeler 1995) examines the sensitivity of a cladogram to the analytical parameters. The data are analysed repeatedly with different parameter sets. Some nodes might be stable over all analyses, while other parts of the cladogram might differ among analyses.

Having obtained a number of possibly differing hypotheses, one is faced with the problem of how to determine the final hypothesis from them. Wheeler (1995) suggested to choose only one of them, using congruence as an external optimality criterion for the decision process. Some measure of congruence is then calculated for each analysis and the cladogram resulting from the parameter set respectively analysis that exhibits maximal congruence is chosen as the best phylogenetic hypothesis.

However, there are a number of problems with choosing a single step matrix and settling on the hypothesis derived from it (see Schulmeister 2003c). I argue in that paper that it is preferable to include only the robust clades in the final hypothesis, which are found in all cladograms resulting from all examined parameter sets. This still does not guarantee that these robust groups are right, but it ensures that the sensitive clades, which are falsified in at least one of the analyses, are being excluded from the final hypothesis. This approach is outlined and justified in detail in Schulmeister (2003c).

The ordering of characters

In Schulmeister (2003b), I discussed how to decide which morphological characters should be treated as additive (ordered).

Maximum parsimony

In Schulmeister (2004), I revisited the issue of the inconsistency of maximum parsimony for the four-taxon case and extended the inconsistency inequality of Felsenstein (1978) to characters with k states. ABSTRACT: "Felsenstein (1978) showed that the method of maximum parsimony can be inconsistent, i.e. lead to an incorrect result with an infinite amount of data. The situation in which this inconsistency occurs is often called the “Felsenstein zone”, the phenomenon also known as “long-branch attraction”. Felsenstein derived a sufficient inconsistency condition from a model for four taxa with only two different parameters for the probability of change on the five branches connecting the four taxa. In the present paper, his approach is used to derive the inconsistency condition of maximum parsimony from the most general model for four taxa, i.e. with five different parameters for the probabilities of change on the five branches and, for the first time, for characters with k states (k = 2, 3, 4, 5, 6, ...). This is used to determine the factors that can cause the inconsistency of maximum parsimony. It is shown that the probability of change on all five branches and the number of character states play a role in causing inconsistency."

References on this page

Felsenstein, J., 1978: Cases in which parsimony or compatibility methods will be positively misleading. Systematic Zoology 27: 401-410.

Schulmeister, S., 2003c: Simultaneous analysis of basal Hymenoptera (Insecta), introducing robust-choice sensitivity analysis. Biological Journal of the Linnean Society 79: 245-275. For additions to and corrections of errors in this paper, see the Errata and Additions sections at the bottom of this page.

Schulmeister, S., and Wheeler, W.C., 2004: Comparative and phylogenetic analysis of developmental sequences. Evolution and Development 6: 50-57.

Schulmeister, S., 2004: Inconsistency of maximum parsimony revisited. Systematic Biology 53: 521-528.

Smith, K. K., 1996: Integration of craniofacial structures during development in mammals. American Zoologist 36: 70-79.

Velhagen, W. A., 1997: Analyzing developmental sequences using sequence units. Systematic Biology 46: 204-210.

Wheeler, W. C., 1995: Sequence alignment, parameter sensitivity and the phylogenetic analysis of molecular data. Systematic Biology 44: 321-331.

Webdesign, Macroxyela photo and HymAToL logo by Susanne Schulmeister