Lab #2: Interpreting phylogenetic trees Objectives: 1. To learn how to interpret phylogenetic trees 2. To compare different trees as alternative hypotheses 3. To interpret characters mapped onto trees as evolutionary hypotheses I BACKGROUND MATERIAL The evolutionary relationships among organisms are always going to be hypothetical: scientists can never go back in time and actually observe the evolution of a group. The formulation of an initial hypothesis about relationships among groups is the process of making observations and constructing a phylogeny, which is a testable hypothesis. This hypothesis can then be tested by the collection of additional data. In this laboratory exercise, you will become familiar with some of the techniques used by systematic biologists to develop hypotheses about the evolutionary history of living things. We have developed several alternative hypotheses proposing the evolutionary relationships among eighteen insect taxa in a program called MacClade. You can use this program to explore the differences among these trees, and then map phenotypic characteristics onto the trees. Using these mappings, you can explore how different trees produce different evolutionary hypotheses for each character. A. THE CHALLENGE OF SYSTEMATICS AND REMOTE INFERENCE (Sys·tem·at·ics n: the study of systems and classification, especially the science of classifying organisms) Both theoretical and practical problems make inferring evolutionary history (a discipline known as systematics) one of the most challenging of the life science disciplines. Like other disciplines in biology, systematics proceeds through the experimental cycle, depending on the construction of hypotheses from observations and the rejection or retention of these hypotheses based on new data. What makes systematics unique is that while these hypotheses address events in the past, often million of years ago, they are based on observations made today. This process is known as remote inference, and you will be employing it in this lab. We shall in all likelihood never see major new characteristics evolving over long periods of time, so we must use observations of both extant organisms and fossils to construct hypotheses concerning the origin of novel phenotypes. Often, within a group of related organisms like the insects, phenotypes can be observed to fall into what appear as logically linked steps: one type is apparently similar to another type. These patterns of similarity can be used to formulate hypotheses that can then be tested against trees of phylogenetic relatedness that are derived from independent data sets. B. TERMINOLOGY Taxon n (plural: taxa): Any of the groups to which organisms are assigned according to the principles of taxonomy, including species, genus, family, order, class, phylum, etc. Phylogenetic tree: A branching diagram in which the branching portrays the hypothesized evolutionary relationships and the sequence of hypothetical ancestors linking observed taxa. Character: An observable aspect of the phenotype of an organism. A character may be genetic (which DNA base is at a particular position in a particular gene sequence), morphological (the shape of a bone or tooth), or behavioral (when behavior is genetically determined, such as in mating behavior or song type). Character state: Across taxa, the detailed structure of a character is likely to vary (if it does not, the character is said to be uninformative). The actual description of a character for a particular taxa is its character state. In systematic analyses, character states are coded numerically, with zero (0) assigned to the character state of the outgroup. Character state matrix: A table of characters where the state of the character in each taxon is coded numerically. Ingroup: The group of organisms in an evolutionary study in which relationships are determined based on the presence of shared characteristics and comparison to the outgroup. There are usually numerous taxa within the ingroup. Outgroup: The taxon least related to any other taxon in an evolutionary study, to which members of the ingroup are compared. There may be one or multiple outgroups in a study. C. PHYLOGENETIC HYPOTHESES Phylogenetic hypotheses are hypotheses about the evolutionary relationships among taxa or among character states. They generally are expressed by biologists in the form of networks (no “ancestor” or root given) or trees (a hypothetical common ancestor or root given). They describe the pattern of speciation that analysis of the data indicate. Trees are built by counting the number of changes (steps from one character state to another) that occur for each of the characters included in the data set. The tree that requires the smallest number of steps over all the characters is assumed to be the best hypothesis. This is a principal known as parsimony, or Occams’ Razor. In science, the first choice hypothesis should be the simplest. It is common in studies of complex groups like “insects” for there to be many alternative trees that are equally parsimonious. Biologists then combine all parsimonious trees, often resulting in a bush. The junctures with multiple branches are referred to as polytomies. In this lab, we will explore the use of hypothetical trees in testing hypotheses concerning character evolution, by posing hypotheses about the evolution of a character and then mapping the character onto a tree to test our hypothesis (where a tree is an assumption about evolutionary relationships). Comparisons among different trees will allow us to examine how our hypothesis fares with different initial assumptions about relationships. Using MacClade First, open FireFox (Mozilla) and choose the Tree of Life bookmark. It should go directly to the insects. Note that on the Tree of Life, the “tree” for the insect groups looks like a garden rake. What does this tell you about scientists’ current understanding of the relationships among insects? Table 1 is a list of the insects that are included in this week’s lab. Some of these are probably familiar to you, and some are likely new. Find each group in the insect boxes in the lab (we do not have examples of the Embiidinae), and on the tree of life. In Table 1, make brief comments or descriptions of the insect types listed. Table 1. In this lab, we’ve chosen the following types of insects for study: asassin bugs cicadas or leaf hoppers aphids lacewings coleoptera: june beetles trichoptera lepidoptera hymenoptera: paper wasp diptera: housefly embiidina zoraptera plecoptera orthoptera phasmida isoptera mantodea earwigs Based upon their general morphology (what they look like), sketch a tree of their evolutionary relationships here: Open MacClade and choose “Phylogenetics lab”. If MacClade opens to show trees, then under “windows”, choose “Data Editor”, and it will switch to show the data matrix. You will see the list of groups from Table 1, and again you can see that for a few groups, a particular member of the group has been chosen. For example, for Diptera (flies), houseflies are specified. Under “Characters”, open two windows: Character List and State Names & Symbols. Tab through the “State Names & Symbols” table until you reach the character “adult mouthparts”. Note that for this trait, the character state will depend upon which fly we choose as an example: the mouthparts on a housefly are quite different from the mouthparts on a mosquito. There are two other groups for which particular species have been chosen, Coleoptera (beetles) and Hymenoptera (wasps, ants and bees). Look at the examples in the collection and at the character states across all characters for these three groups and be sure you understand why in this lab exercise we had to choose particular types for these groups. Tab through the State Names and Symbols list again and become familiar with the different states. What exactly does each description mean, and how is it related to the biology of the organism? Look at the list for character 1, metamorphosis, which has two character states (presence or absence of metamorphosis). On your tree, indicate which organisms have which state. Now, trace backwards on the tree, assigning character state 0 (no metamorphosis) or 1 (metamorphosis) to each of the intermediate branches. For example, if you have two taxa that are sisters (share an ancestor), one with character state 0 and one with 1, the immediate ancestor has character state 0/1 (in other words, you can’t know). At the next branch point, compare 0/1 to the character state in the next-closest relative; we’ll say that it is “1”. Now you know that the ancestral state of the ambiguous branch was “1”, and that the common ancestor to all three of these taxa is also assigned “1”. You can work your way down the entire tree in this fashion, assigning “1”, “0” or “1/0” to each ancestor. Now we are ready to look at some trees in MacClade. Under “Windows”, select Tree Window. Your Characters and State Names & Symbols windows should stay open; if they don’t, reopen them (the data editor window will disappear). The tree should be black; if it isn’t go to the “Trace” pull-down menu and turn Trace off. A mini-window will also open inside the tree window. It gives you the tree name and its length (the number of steps or changes in the entire tree: shorter is more parsimonious). In this window, you can scroll from tree to tree. Scroll until a tree named “Polytomy” is displayed. This is the same tree (drawn differently) as the rake on the Tree of Life web page. Scroll through all the trees on the list. Think about the organisms listed, and consider whether the hypothesized relationships among them are consistent with the general body shape that they have. Which tree is most similar to the tree you drew in conjunction with Table 1? Which tree is most parsimonious (shortest treelength)? Which branches would you have to move to make your tree look more like the most parsimonious tree in the MacClade file? Tracing character evolution: Return to the polytomy tree, and under the “Trace” pull-down menu, turn Trace on by selecting “Trace Character”. A second miniwindow will appear which tells you which character you are displaying the trace of. Scroll until you are displaying “metamorphosis”. The tree now displays an hypothesized pattern of evolution of the character “metamorphosis”, the window lists the title of the character that is being displayed and the number of evolutionary steps (changes from one state to another) that are represented by that particular tree. Each character state has a color, and “equivocal” is hatched. When a character state is “equivocal”, it means that in this program, two or more character states are possible for that particular branch in that particular tree. Are any branches “equivocal” for metamorphosis? The tree mappings also present hypothetical ancestral states for a character: the phenotype that a hypothetical ancestor would have. What is the ancestral state for metamorphosis? Scroll through the other characters. How do the ancestral states compare to the character states for the outgroup? List all characters with the identified ancestral state for the polytomy tree in Table 2 (write equivocal if there is no ancestral state): Table 2: Ancestral Character states according to the Polytomy hypothesis. Character Ancestral State (polytomy) number of steps It should be clear that the polytomy is not very informative for answering questions about the pattern of evolution of characters. Return to the tracing of metamorphosis, and select another tree from the tree list. Note in column 2 of Table 3 which tree number you have chosen, then fill out the ancestral states for each character and the number of steps for the evolution of each. Table 3: Character Ancestral State (tree _____) number of steps How does the increased resolution improve our understanding of the evolution of these characters? Character tracing: Now that you are familiar with the function of the character tracing, you will test hypotheses about character evolution. To illustrate how this is done, we will do one character, social structure, together. Social structure. It is commonly hypothesized that extended parental care is a pre-condition for the evolution of complex social structure in animals. Therefore, we would hypothesize that the pathway for the evolution of social behavior would take the path: none ––> extended parental care ––> extended families ––> social. We can test this hypothesis by mapping social behavior onto a tree. In tree 1, the ancestral state is “no social structure”. However, do the relationships among groups with parental care, extended families, and sociality show the evolutionary pattern we’ve hypothesized? How many steps are in this mapping? Move through the different trees. Examine how the different tree structures affect the hypothesis on the evolution of social structure. As tests of the above hypothesis on the evolution of social structure, these trees range from complete rejection to luke-warm support. Which tree supplies the best support for the hypothesis? It is very important to note that when testing hypotheses about the evolution of characters, biologists ought not to choose among trees based upon the character mapping. The tree is built using characters that are not the subject of current study. Indeed, under the best circumstances, the characters used to build the tree will have no obvious biological link to the characters being mapped. Commonly these days, the characters used are gene sequences and protein structure. Exercises: I. Choose one character from the following list: Wing folding Silk spinning Wings Examine the character states for your chosen character, and if needed look at either the insects on display or on the Tree of Life. Below, write an evolutionary hypothesis concerning the pathway of change of these characters, noting the original (ancestral) state and the transitions that seem logical to you. Go to tree 1, then choose your character from the character list. Does this mapping support or reject your hypothesis? Note that if key branches on the tree are labeled “equivocal”, that tree is uninformative for your hypothesis. Scroll through the fully resolved trees. Describe in Table 4 below how each tree affects your evaluation of the hypothesized evolutionary pathway for the character you’ve chosen: Table 4 Tree number of steps evaluation of hypothesis II. Choose two characters from the following list, and repeat the exercise: Adult mouthparts food (adult) food (larvae) mating behavior Table 5: Character_______________ Tree number of steps evaluation of hypothesis Table 6: Character______________ Tree number of steps evaluation of hypothesis Draw one of the trees below, and hand-map onto the tree the points of change for each character you’ve chosen, as you did for metamorphosis at the beginning of the lab. Count the number of changes, including reversals. Be sure you understand how the mapping is done, and how this tree is identical to the tree produced by MacClade.