Integrative Biology 200A "PRINCIPLES OF PHYLOGENETICS" Spring 2006 University of California, Berkeley B.D. Mishler Jan. 31, 2006. Character Analysis A. Introduction to the logic of the data matrix: The process of phylogenetic analysis per se inherently consists of two phases: first a data matrix is assembled, then a phylogenetic tree is inferred from that matrix. There is obviously some feedback between these two phases, yet they remain logically distinct parts of the overall process. One could easily argue that the first phase of phylogenetic analysis is the most important; the tree is basically just a re-representation of the data matrix with no value added. Paradoxically, despite the logical preeminence of data matrix construction in phylogenetic analysis, by far the largest effort in phylogenetic theory has been directed at the second phase of analysis, the question of how to turn a data matrix into a tree. At stake are each of the logical elements of the data matrix: the rows (what are the terminal units or OTUs?), the columns (what are the characters?), and the individual entries (what are the character states?). The elements of a data matrix (note the interlocking definitions): OTU = group of semaphoronts that can’t be subdivided given current character data Character = an apparently homologous feature, independently varying among OTUs Character-state = a discrete condition within a character, potentially a phylogenetic marker B. What is an OTU? These are represented by rows in the data matrix. People are usually cavalier about what their terminal branches represent. One often sees species or other taxon names, or even geographic designations of populations, attached to terminal branches of published trees without explanation. Larger-scale units might indeed be a well-justified OTU, but they need to be justified by preliminary analyses, never assumed a priori. Taxa or populations are never the fundamental things from which phylogenies are actually built. Not even individuals are the OTUs -- so what is the fundamental OTU? As was carefully elaborated by Hennig (1966), the fundamental terminal entity in phylogenetics is the semaphoront, an instantaneous time slice of an individual organism at some point in its ontogeny. A tube of extracted DNA and its associated museum voucher specimen, photos, data, etc. —a semaphoront— should be considered the ultimate unit. An OTU is an agglomeration of semaphoronts, that are not divisible by the characters currently known. Hence, the interrelationship between the concept of OTU and character. [More later in the class when we cover species concepts.]C. What is a Character? The basic ontological stance taken here is: A taxonomic character (=putative taxic homology) is a piece of evidence for the existence of a monophyletic group. The central epistemological problem of systematic research is thus how to recognize, distinguish, and "define" taxonomic characters precisely. The selection of characters can be viewed as a sort of a priori weighting (Neff, 1986), but realize that this is quite a different issue than a posteriori weighting (more on that in a later lecture). Epistemologically, a good taxonomic character is one that shows convincing potential homology across the OTU’s being considered, and shows greater variation among OTU's than within. This variation must be heritable and independent of other characters, i.e., not genetically correlated with other characters in a specific evolutionary sense. Note that there are other meanings of "correlation", some of which (such as phylogenetic congruence) do not disqualify characters from counting as independent. Note also that this view of taxonomic characters requires that each be a system of at least two discrete transformational homologs, or character states (more below). Note that this is a restricted usage of the term "character," derived from the ontology of phylogenetic systematics. For other purposes, as in functional/evolutionary studies, numerical phenetic comparisons, or identification, less strict usages are applied. D. What is a character state? The ontological view of taxonomic characters discussed above requires that each be a system of at least two discrete transformational homologs, or character states. Epistemologically, the distinction of character states is a issue involving patterns of variation among OTUs. How can we divide a "quantitative" character into states? There are two extremes of opinion: gap coding (e.g. Archie, 1985, Syst. Zool. 34: 326-345) vs. elimination (="range coding" -- e.g. Pimentel and Riggins, 1987). Both are too extreme. A reasonable intermediate approach can be outlined as follows (see also Mishler & De Luna, 1991): Given three OTU's (A,B,C) already tentatively distinguished on the basis of other characters, and the concept of character defended above, the question is: does a new quantitative character (one for which A, B & C have different means, but overlapping ranges, with A the largest, C the smallest, and B intermediate) provide evidence for the existence of monophyletic groups within this collection of OTU's? In more precise terms: is there a statistically significant association of the quantitative feature with the a priori discrete groups? An even more precise way to form the question is: can we reject the null hypothesis that the means of each OTU for this feature came from the same underlying parametric distribution? This is a standard type of question in science, and there is a generally applicable body of statistical methods known as analysis of variance to deal with such questions. In particular for our purposes we are interested period of shared historyin the various multiple comparison tests designed to discover "which pairs of means are different from each other, and whether the means can be divided into groups that are significantly different from each other." (Sokal and Rohlf). [see handout on Tortula] An all too common problem, however, with the results of multiple comparison tests is a situation as follows: statistically _____________ homogeneous _______________ subsets: A B C What to do? It may be that B is heterogenous, and if broken up into two OTU's might resolve the situation. This is not valid if the overlapping variation occurs within individual organisms or
View Full Document
Unlocking...