**EvoWiki is now a project of the RationalMedia Foundation.We are moving all content to RationalWiki.See the EvoWiki project page for details!**

# Systematics

### From EvoWiki

**Systematics** is the scientific study of the kinds and diversity of organisms and of any and all relationships among them. (After Simpson 1961).

Methodological approaches for systematic studies vary widely, depending in part on the type of data being examined, as well as (to some extent) on the philosophical tastes of the researcher. The overall goal of any method in systematics is to find the dendrogram that best accounts for the data and that presumably reflects the pattern of descent within a clade.

Generally, systematics methods can be classified as either **algorithm-based** or **criterion-based**. Algorithm-based methods are those that apply a procedure (usually iteratively) to a dataset to build up a dendrogram. An example of an algorithmic method is neighbor-joining, in which the most similar taxa are successively joined.

Criterion-based methods are those that calculate a statistic for each given dendrogram, and then attempt to find that dendrogram with the most extreme value of that statistic. In the case of parsimony-based methods, the criterion is the number of 'steps', or character-state changes, and the objective is to minimize this value, i.e., to find the tree with the fewest number of steps. An example is cladistics, which seeks the most parsimonious distribution of synapomorphies.

Another class of criterion-based methods are those that depend on a model of evolution. The best example of this is the method of maximum likelihood. With this method, the objective criterion is the statistical function likelihood. Very generally, the likelihood of a tree under a particular model is proportional to the probability that that tree generated the data, given the model.

If the number of taxa in the dataset is small enough, finding the maximum likelihood or maximum parsimony tree can be accomplished by exhaustion: every possible tree is examined, and the best (by the criterion) is kept. (This assumes, of course, that there is a single "best" tree, which is rarely true.) For larger datasets, the entirety of "tree-space" cannot be examined, so some sort of heuristic searching algorithm must be employed.

## References

- Hillis, D.M., Moritz, C. & Mable, B.
*Molecular Systematics*. Sinauer. - Hull, D.L. 1988.
*Science as a Process*. University of Chicago Press.