19 Feb: Methods in plant systematics

Our working definition of a phylogeny is: an hypothesis of the evolutionary relationships among organisms. But "organisms" is really too restrictive. Phylogenies can infer the relationships of individual genes, biochemical pathways, languages, basically anything that evolves.

What can you do with a phylogeny? See phylogeny phlow chart handout - you can use them to study: 1) taxonomy and classification; 2) diversification patterns and biogeography; 3) reproductive ecology, adaptation, speciation/extinction, conservation; 4) crop improvement, disease transmission, drug prospecting; 5) genomics, molecular evolution, development.

Make sure you are comfortable with the definitions of: systematics, classification, taxonomy, convergence, ancestral, derived.

These bear repeating (and see figure on page 15 of Zomlefer):
A monophyletic group is composed of an ancestral taxon and ALL of its descendants. Natural classification schemes strive to name monophyletic groups.

A paraphyletic group is composed of an ancestral taxon and SOME but NOT ALL of its descendants. (Some descendant taxa may have diverged significantly, e.g. birds from within reptiles).

A polyphyletic group is composed of organisms that DO NOT share a common history (they have origins along 2 independent lineages). Not a natural group.

Phylogenetic classification attempts to give names only to monophyletic groups.

Beware: many traditionally recognized plant family names have been shown to be paraphyletic. In modern texts (including Zomlefer, your course text), they have been recircumscribed to recognize only monophyletic groups but in many floras and keys, the original family names persist. An example is the Ericaceae, which is monophyletic only when previously recognized families such as the Pyrolaceae and Empetraceae are included within it. Hultén considers Pyrolaceae and Empetraceae distinct families.

Similarities used to identify groups and relationships can reflect 3 different types of ancestor/descendant relationships:

synapomorphy: a derived character, shared by two or more taxa.
Synapomorphies are evidence of evolutionary relationship and can be used to define monophyletic groups (clades). By definition, a synapomorphy is unique to the clade it is in. Examples would be feathers on birds, these evolved at the time birds first appeared. In plants an example would be the enclosed seed of angiosperms. Synapomorphies are homologies and reflect monophyletic groups.

Symplesiomorphy: a shared, ancestral character. Can't be used to define a monophyletic group (e.g. tricolpate pollen is symplesiomorphic within eudicots - individual eudicot families cannot delineated based on whether they have this feature). Examples would be the keratin scales on reptiles, these were transformed to feathers when birds branched off. In plants, the naked seed of "gymnosperms" was transformed into an enclosed seed when angiosperms evolved. Symplesiomorphies are also homologies but reflect paraphyletic groups.

Convergent similarities: are due to evolution under similar environmental conditions or selective pressures, NOT due to a common ancestor. Examples would be wings on bats, birds and insects, each represents how these groups solved the same problem of harnessing aerodynamics. In plants, an example might be succulence in several diverse plant families living in xeric conditions. Convergent similarities are called homoplasies and represent polyphyletic groups.

Evolution involves the separation of lineages (ancestor-descendant sequences of populations), and can be visualized by examining these processes of change on a phylogenetic tree.

To construct or infer (remember, it's an hypothesis) a phylogeny you need to:
1) define your ingroup (the group of organisms you would like to study. It's common to use a phylogeny to test the monophyly of an ingroup)

2) identify characters (any attribute of an organism that can be measured - morphological structure, chemical, anatomical feature, DNA sequence).

3) score character states (the alternate values of a character - flower color red or white, stem condition woody or herbaceous, DNA base at a particular position A, T, G or C, presence or absence of a structure or compound). Continuous character states like the length of a leaf can also be used, but are often problematic.

Characters used in plant systematics have changed through time. Some sources of character data (oldest to most recent) are: 1)morphology, 2)anatomy (cellular organization of plants), 3)embryology and development, 4)cytology (chromosome number), 5)secondary chemistry (chemical compounds produced by plants not used in primary metabolic processes like photosynthesisl. Examples are terpenoids, flavonoids), 6)immunology (method to estimate protein similarity using antibodies), 7)protein electrophoresis (method for identifying specific enzyme allelic forms), 8)DNA (comparison of restriction fragment length, or direct DNA sequence alignment and comparison), 9)genomics (newest way to compare organisms, entire genome is compared. Not yet really tractable)

With this information, you can construct a character or data matrix and from the matrix you can create a network of relationships.

In order to add a time (or evolutionary direction), you need to determine which characters are ancestral and which are derived. To do this you need to polarize the characters in the matrix. Polarity assessment can be accomplished by:

a)fossil record (oldest is primitive), b)simple to complex (evolutionary trends tend to be parallel between groups, c)correlation (primitive characters often tend to occur together in organisms, d)ontogeny (developmentally early stages are primitive), e)the method that is applicable most uniformly and has the fewest problems is outgroup (a taxon from a lineage that separated from the ingroup BEFORE the lineages within the ingroup diversified) comparison.

The number of possible rooted tree topologies increases exponentially as you increase taxa, so a criterion (and method) to find the "best" trees is required. There are several methods in common use based on the principle of parsimony (the principle that the explanation requiring the least number of changes is preferred), evolutionary distance, or statistical likelihood. References available for those interested. There are also several ways in which you can test the statistical robustness of phylogenetic trees.

Trees have branches, nodes (where branches connect), tips (the taxa at the ends of the branches), and a root, which implies directionality and time. Branches can be rotated around their nodes which does not affect relationships (but can be visually confusing).

Characters of any type can be mapped onto phylogenetic trees and trees can be used to infer the number of origins or the pathway of evolutionary change a character has undergone through time.

To go from phylogenies to classifications:

Again, phylogenetic classification attempts to give names only to monophyletic groups.

Only monophyletic groups are named, but not all monophyletic groups are given formal names.

Criteria for naming plant groups (for classification purposes, NOT the actual taxonomic naming of new species, for which a very specific code of rules exists):

There is no codified set of rules for deciding which monophyletic groups to name, but there are some informal criteria that are generally followed:

1) Strength of evidence. Only groups linked by many shared derived characters should be named. A clade should have many features that allow it to be distinguished from other clades or diagnosed.

2) Obvious morphological synapomorphies. Classifications should be useful to biologists who aren't trained in systematics, so naming a monophyletic group based only on molecular characters is not helpful to field biologists and should be avoided.

3) Size of group. Classification is directed toward information retrieval. Extremely large groups are more tractable if divided into smaller units, but large groups should be divided only if there are reasonably well-defined differences.

4) Nomenclatural stability. Classification is essentially a means of communication - a working vocabulary of biological diversity. If names are constantly being changed, the most important function of a classification is lost. Diagnosable, monophyletic groups that have been named in the past should continue to be named in new classification schemes.

Ranks
Monophyletic taxa represent real groups that exist in nature as a result of the process of evolution (shared history). The categorical ranks of Linnaean classification are mental constructs ­ there are no special criteria that determine whether a particular group should be considered a phylum, class or order. Ranks are arbitrary, and nomenclatural stability is an important factor in catorical sequencing.

There is a growing school of thought among systematists that since ranks have no real biological meaning, they should be abandoned in favor of unranked classifications.