Phylogenetic Investigator Version 2.0.1 User`s Manual

Phylogenetic Investigator Version 2.0.1 User`s Manual
Phylogenetic Investigator
Version 2.0.1
User’s Manual
Steven D. Brewer
Robert Hafner
University of Massachusetts
Western Michigan University
A BioQUEST Library VII Online module published by the BioQUEST Curriculum Consortium
The BioQUEST Curriculum Consortium (1986) actively supports educators interested in the
reform of undergraduate biology and engages in the collaborative development of curricula.
We encourage the use of simulations, databases, and tools to construct learning environments
where students are able to engage in activities like those of practicing scientists.
Email: [email protected]
Website: http://bioquest.org
Editorial Staff
Editor:
Managing Editor:
Associate Editors:
John R. Jungck
Ethel D. Stanley
Sam Donovan
Stephen Everse
Marion Fass
Margaret Waterman
Ethel D. Stanley
Online Editor:
Amanda Everse
Editorial Assistant: Sue Risseeuw
Beloit College
Beloit College, BioQUEST Curriculum Consortium
University of Pittsburgh
University of Vermont
Beloit College
Southeast Missouri State University
Beloit College, BioQUEST Curriculum Consortium
Beloit College, BioQUEST Curriculum Consortium
Beloit College, BioQUEST Curriculum Consortium
Editorial Board
Ken Brown University of Technology, Sydney, AU
Joyce Cadwallader St Mary of the Woods College
Eloise Carter Oxford College
Angelo Collins Knowles Science Teaching Foundation
Terry L. Derting Murray State University
Roscoe Giles Boston University
Louis Gross University of Tennessee-Knoxville
Yaffa Grossman Beloit College
Raquel Holmes Boston University
Stacey Kiser Lane Community College
Peter Lockhart Massey University, NZ
Ed Louis The University of Nottingham, UK
Claudia Neuhauser University of Minnesota
Patti Soderberg Conserve School
Rama Viswanathan Beloit College
Linda Weinland Edison College
Anton Weisstein Truman University
Richard Wilson (Emeritus) Rockhurst College
William Wimsatt University of Chicago
Copyright © 1993 -2006 by Steven D. Brewer and Robert Hafner
Copyright, Trademark, and License Acknowledgments
Portions of the BioQUEST Library are copyrighted by Annenberg/CPB, Apple Computer Inc., Beloit College, Claris
Corporation, Microsoft Corporation, and the authors of individually titled modules. All rights reserved. System 6,
System 7, System 8, Mac OS 8, Finder, and SimpleText are trademarks of Apple Computer, Incorporated. HyperCard
and HyperTalk, MultiFinder, QuickTime, Apple, Mac, Macintosh, Power Macintosh, LaserWriter, ImageWriter, and the
Apple logo are registered trademarks of Apple Computer, Incorporated. Claris and HyperCard Player 2.1 are
registered trademarks of Claris Corporation. Extend is a trademark of Imagine That, Incorporated. Adobe, Acrobat, and
PageMaker are trademarks of Adobe Systems Incorporated. Microsoft, Windows, MS-DOS, and Windows NT are
either registered trademarks or trademarks of Microsoft Corporation. Helvetica, Times, and Palatino are registered
trademarks of Linotype-Hell. The BioQUEST Library and BioQUEST Curriculum Consortium are trademarks of Beloit
College. Each BioQUEST module is a trademark of its respective institutions/authors. All other company and product
names are trademarks or registered trademarks of their respective owners. Portions of some modules' software were
created using Extender GrafPak™ by Invention Software Corporation. Some modules' software use the BioQUEST
Toolkit licensed from Project BioQUEST.
ABOUT PHYLOGENETIC INVESTIGATOR
Evolution, the central theme in biology, takes on added meaning for
students when they can explore the construction and interpretation of
evolutionary models. Phylogenetic Investigator (PI) facilitates creative problemsolving in phylogenetic inference for teaching and learning evolutionary biology.
Users can identify characters and states, polarize characters, and engage in
directed-search phylogenetic tree construction. PI also allows the user to (1)
make inferences and represent them one step at a time, (2) vary
representational features of their trees (such as angle of divergence and time
between speciation events), (3) create reticulate tree patterns, and (4) view all
of the character transformations at one time. In addition, PI can generate
plausible data stochastically for modeling and practicing tree construction.
Phylogenetic Investigator was developed with support from the
Department of Science Studies at Western Michigan University in Kalamazoo,
Michigan . PI was created using SuperCard®. Portions ©1989-1994 Allegiant
Technologies, Inc.
TABLE OF CONTENTS
A PRIMER ON PHYLOGENETIC SYSTEMATICS.........................................3
Introduction.............................................................................3
Phylogenetic Trees...................................................................4
A Brief History of Systematics ..................................................6
A METHODOLOGY OF PHYLOGENETIC INFERENCE .................................8
Assumptions............................................................................8
Phases of Phylogenetic Inference ..............................................11
Selection of Ingroup and Outgroup ............................................11
Determination of Characters and States.....................................11
Assignment of Polarity .............................................................12
Outgroup method ....................................................................12
Paleontological method ............................................................12
In-group method ......................................................................13
Tree Construction ....................................................................13
AN EXAMPLE PROBLEM USING PI.........................................................16
PHYLOGENETIC INVESTIGATOR REFERENCE MANUAL ............................31
Windows .................................................................................31
Chars & States ........................................................................31
Small configuration ..................................................................32
Large configuration ..................................................................33
Data Matrix..............................................................................33
Phylogenetic Tree ....................................................................34
Menus .....................................................................................35
Apple ......................................................................................35
File .........................................................................................36
Edit.........................................................................................37
Actions ...................................................................................37
Problems .................................................................................38
Set-Up Problem........................................................................38
Model Problems........................................................................39
Practice Problems ....................................................................39
Windows .................................................................................39
OTHER SOFTWARE FOR PHYLOGENETIC ANALYSIS ...............................40
SUGGESTED READINGS ......................................................................41
BIBLIOGRAPHY ..................................................................................42
APPENDIX A -- MODEL PROBLEMS .......................................................44
APPENDIX B -- INSECT WING DATA SOURCE.........................................52
Phylogenetic Investigator 2
A PRIMER ON PHYLOGENETIC SYSTEMATICS
Introduction
What is phylogenetic systematics and why do people do it?
Each 'living thing' (or organism) is unique. Descended from some ancestor
or ancestors and potential progenitors of offspring, organisms exist in
populations of related organisms (species). Humans everywhere have named
the species around them and evaluated the properties of each. Knowing
whether a species was edible, medicinal, or poisonous could mean the difference
between life and death. One of the fundamental aims of biology has been to
create a nomenclature, or system of terms, that could systematically
encompass the natural world.
It is axiomatic that species fall into natural kinds (See "A Quahog is a
Quahog" in The Panda's Thumb Gould (1980)). Birds, although there are many
different species, share features that appear to set them apart from all other
kinds of living things. Similarly, these natural kinds seem to have some kind of
hierarchical organization that can be represented by a taxonomy with species as
the most basic taxon, or grouping, which can be placed within more and more
inclusive taxa. A Red-winged Blackbird is one kind of blackbird which is one kind
of perching bird which is one kind of bird which is one kind of the animals with
backbones which is one kind of animal, and so on. Charles Darwin put forward a
coherent explanation for this phenomenon that has come to be widely
accepted. The theory of evolution proposes that living things are somehow
related through ancestral/descendant relationships and that very similar things
are more closely related than less similar things. Before a theory of evolution,
taxa were usually based on the principle of overall similarity. The goal of
phylogenetic systematics is the construction of a taxonomy based not on
similarity, but on evolutionary relationship or genealogy.
The ability to describe how species are related has transformed how
scientists understand evolution, systematics, and biogeography. Recently an
issue of Bioscience was devoted to phylogenetic systematics (Simpson and
Cracraft, 1995). Phylogenetic systematics, as a means to interpret the
properties, activities, and distributions of species and groups of species, is
illustrated from a variety of perspectives: biodiversity (Savage, 1995),
agriculture (Miller and Rossman, 1995), ecology and behavior (Brooks et al,
1995), the study of organismal form and function (Lauder et al, 1995), and
public health (Davis, 1995). In each of these examples, the ability to recognize
the underlying relationships among species allows insight into the processes
that have led to current conditions and makes it possible to predict future
trends.
Phylogenetic Investigator 3
Phylogenetic Trees
What do they look like and what do all those things mean?
This section provides a brief description of phylogenetic trees, as they are
conceptualized in Phylogenetic Investigator. Some of the concepts presented
here are described at greater length elsewhere in the text.
A phylogenetic tree is a diagram (Fig. 1) with time on the Y axis and
evolutionary change (in PI this is assumed to be morphological change) on the X
axis that illustrates a hypothesis of evolutionary relationships and the sequence
of evolutionary events that gave rise to some group of taxa of interest (termed
'the ingroup'). In PI, phylogenetic trees are constructed of three kinds of pieces:
nodes, links, and transitions.
Nodes represent taxa, for example species. Designations for nodes can
have the prefix R, F, or P. Nodes that correspond to the observed taxa that are
being studied, are numbered and have a letter prefix that is either R for Recent
or F for Fossil. The ingroup in Figure 1 consists of R80, R86, R84 and R82. F98
is a fossil taxon from which the ingroup is believed to have descended. During
tree construction, common ancestors of taxa are postulated to have existed in
order to explain the data. Each of these nodes has a letter (e.g. A, B, C, etc.)
with the prefix P (for Postulated).
Links connect nodes and represent hypothesized ancestor/descendant
relationships between taxa. The slope of a link indicates the rate of
morphological change: vertical lines indicate no change over time and the more
a line tends to the horizontal, the more rapidly change is perceived as having
taken place.
Transitions appear on links and represent the point at which evolutionary
changes are believed to have occurred. Each transition represents some feature
(character) of the taxa which has been numbered and described as having two
conditions (states). One state is considered ancestral and is coded with a "0".
The evolutionarily novel (or derived) state is coded with a "1". A transition
shows the point where a character changes from "0" to "1" or from "1" to "0".
Coded characters and states are organized by taxa in an associated data matrix.
Phylogenetic trees are just one type of a kind of branching diagram that
appears often in biology. Other branching diagrams in biology include
genealogies, that show relationships among individuals, and fate maps, that
show how cells become canalized during the early stages of development. Both
of these diagrams seeks to represent the systems of relationships that result
from selective and reproductive processes at different hierarchical levels in
biology (phylogeny at the level of species,
Phylogenetic Investigator 4
Time
0
R80
R86
R84
R82
10
PB
PA
20
2 O>l
1 O>l
30
PC
3 O>l
40
50
F98
Morphological Change
3 Steps
1
R80 l
R82 O
R86 l
R84 O
T F98 O
Characters
2
O
l
O
l
O
3
l
l
l
l
O
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20
a
x
a
Problem: Synapomorphy 3
Figure 1.
A phylogenetic tree as constructed using Phylogenetic
Investigator.
Phylogenetic Investigator 5
genealogy at the level of individuals, and fate map at the level of cells). At the
evolutionary level, these processes are microevolution (which
causes lines to have a slope), speciation (which causes lines to branch), and
extinction (which causes some taxa to leave no descendants).
Phylogenetic trees typically have dichotomous branching patterns, but
trichotomies and even polytomies are possible. Each taxon is usually assumed
to be derived from a single ancestral species, but using PI it is possible to
create links to more than one ancestral species. These reticulating tree
structures are occasionally used to illustrate hypotheses of interspecific genetic
transfer (for example, hybridization).
A Brief History of Systematics
Traditional Linnaean classification still dominates how systematics is
taught in most introductory biology texts. Linnaeus viewed species as unique
and unchanging types or natural kinds. Each natural kind, according to Linnaeus,
had particular morphological features that defined it. By describing those
features systematically as taxonomic characters (a character being any
attribute of an organism or group by which it may differ from another organism
or group), each kind could distinguished from every other kind. Darwin's theory
of evolution called for species to be historical entities which could change over
time, produce new species, and go extinct. Systematics as a discipline has still
not recovered from the impact of evolutionary theory and continues to be
transformed today.
Systematics has become divided into two main schools of thought based
primarily on different conceptions of the taxonomic goal (For a review see
Ridley, 1986). Phenetic systematics seeks to represent a hierarchy based on
the similarity of living things while phylogenetic systematics seeks to represent
the hierarchy of evolutionary change. These forms of classification often result
in similar, but different groupings. Phylogenetic inference seeks to define sets
of species (taxa) which are all descended from one ancestral species
(monophyletic). An incomplete set of descendant species is paraphyletic while a
set which contains unrelated species is polyphyletic (Fig. 2). Phenetic
classifications have been criticized because they sometimes group organisms
that appear similar due to convergent evolution, but which are actually only
very distantly related (resulting in polyphyletic groupings). They also sometimes
fail to group things which are evolutionary related, but which have diverged
greatly from one another (resulting in paraphyletic groupings) .
Although both phylogenetic and phenetic systematics seek to define
groups based on shared similar characters, phylogenetic systematics makes a
fundamentally different inference about the nature of some shared characters.
Whereas phenetic classification treats all characters equally, phylogenetic
classification is based solely on characters that are believed to demonstrate
shared ancestry.
Phylogenetic Investigator 6
Time
0
R01
R02
R04
R05
R03
R06
R07
10
PF
PE
20
PC
PD
30
PG
40
PB
50
PA
Morphological Change
Figure 2.
The placement of taxon R03 illustrates a paraphyletic
grouping of {R01, R02 and R04} and a polyphyletic
grouping {R05, R03, R06 and R07}. A group composed of
all of the recent taxa {R01, R02, R03, R04, R05, R06,
R07} is monophyletic.
Organisms share characters either because they are the result of shared
ancestry (homology) or because they have evolved convergently in separate
organisms (analogy). Only characters showing homology are useful for inferring
phylogenetic relationships. In turn, homologous characters can be shared either
because a character is generally ancestral or because it is modified from the
ancestral. Ancestral characters may be retained by any combination of taxa
regardless of phylogenetic relationship, but derived characters will be shared
only by descendants of the ancestral species in which the character evolved.
Therefore, only shared, homologous characters in the derived condition are
useful for inferring phylogenetic relationships.
Phylogenetic Investigator 7
A METHODOLOGY OF PHYLOGENETIC INFERENCE
Should I draw phylogenetic trees and how do I do it?
Assumptions
Phylogenetic trees are hypotheses about how taxa are related to one
another. Constructing phylogenetic trees requires a number of critical
assumptions: (1) that all species in the ingroup, are descended from a single
common ancestor, (2) that shared similarities among species are the result of
sharing more recent common ancestors, (3) that ancestral and derived states
of characters can be determined, and (4) that some form of character
congruence indicates the most probable path of evolutionary relationship.
Phylogenetic inference will yield accurate results to the extent that these
assumptions are warranted. The reader should note that what is presented here
is a general account of phylogenetic inference or what is sometimes termed
Hennigian argumentation. Some recent forms of phylogenetic inference allow
rejection or suspension of some of these assumptions.
The first assumption is an assumption of evolutionary process.
Ancestral/descendant relationships, resulting from evolutionary processes, tie
the diversity of living and fossil organisms together into a meaningful
framework. Without this assumption, there would be no reason for supposing
that there was any kind of underlying relationship among living things and
phylogenetic inference would be meaningless. One could go through the
mechanics of making groups based on shared derived characters, but there
would be no coherent reason for doing so. (In fact, one school of systematics,
which has come to be called transformed cladistics, has separated from the
phylogenetic school arguing that the existence of patterns of character
congruence, irrespective of models of evolutionary process, can serve as the
raison d' etre for a systematic methodology. See Ridley, 1986 for a review.) On
the other hand, the fact that phylogenetic inference appears to yield meaningful
results is one of the pieces of evidence that has been used as support for the
theory of evolution.
The second assumption deals with whether or not it is reasonable to
postulate the links and common ancestors that will be used to construct a
phylogenetic hypothesis. It is easy to imagine cases where this assumption
would not be warranted and would result in a misleading analysis. Imagine the
case of a species distributed over a continent which is subsequently inundated
in a single event resulting in 5 islands with reproductively isolated populations. If
the disjoint populations eventually evolved into 5 different species, one could
deduce that any derived character states shared by these species could not be
the result of recent common ancestors (Fig. 3). (Note: This issue is somewhat
more complex than indicated here because, although there can be no common
ancestors among populations after the inundation, some derived characters
Phylogenetic Investigator 8
may have had their origin prior to the separation of the populations and only
been driven to fixation afterwards.) In this case the true phylogeny (Fig. 3) has
only convergent characters. Every seemingly shared character must have
Time
0
R01
R02
R03
R04
R05
10
1 O>l
1 O>l 1 O>l 2 O>l
1 O>l 2 O>l 2 O>l 3 O>l
3 O>l 4 O>l
20
1
2
3
4
5
O>l
O>l
O>l
O>l
O>l
30
40
PA
50
Morphological Change
Characters
15 Steps
1 2 3 4 5
R01
R02
R03
R04
T R05
l
l
l
l
l
O
l
l
l
l
O
O
l
l
l
O
O
O
l
l
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
O
O
O
O
l
a
x
a
Problem: Island Problem
Figure 3.
The "true" phylogeny in which 5 species are descended
from a single common ancestor.
been independently acquired in each population because no more recent shared
ancestors are possible. Methods of phylogenetic inference, however, would still
yield a tree that explained all shared derived characters using shared common
ancestors. In this case that assumption is unwarranted and the resulting
phylogeny (Fig. 4) would be incorrect.
The third assumption deals with the determination of states of
characters. If we cannot tell which characters are derived, then we cannot make
groups on the basis of shared derived characters. Several techniques (e.g.
outgroup, paleontological, and ingroup methods) are available for making
determinations of states of characters and, although none are perfect, each can
be evaluated to consider whether or not it can be counted on to provide
meaningful results (Stuessy and Crisi, 1984). Futhermore, often several
methods can be used and their results used to corroborate each other.
The last assumption deals with the issue of characters that suggest
contradictory histories of descent. This can occur either because ancestral
character states have been mistaken for derived states, or because of
homoplasy (convergent evolution): either parallel appearance of a character in
Phylogenetic Investigator 9
the derived state or reversal of a character back to the ancestral state. In some
cases, further study of the taxa themselves can illuminate the source of the
conflict. A closer look at the taxa may show that two structures which appeared
homologous are, in fact, substantively different. If further study does not
diagnose the source of the conflict,
Time
0
R01
R02
R03
R04
R05
5 O>l
10
PE
4 O>l
20
PD
3 O>l
30
PC
2 O>l
40
PB
1 O>l
50
PA
Morphological Change
Characters
5 Steps
R01
R02
R03
R04
T R05
1
l
l
l
l
l
2 3 4 5
O O O O
l O O O
l l O O
l l l O
l l l l
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20
a
x
a
Problem: Island Problem
Figure 4.
A phylogenetic tree representing 5 species descended from
a single common ancestor through a nested series of more
recent common ancestors.
statistical methods can be used to provide a basis for determining which of the
possible trees should be preferred (Harvey and Pagel, 1991). The most
common criterion has been termed parsimony and refers to selecting the tree
that requires the fewest character state changes to explain the data. This
criterion is based on the assumption that evolutionary events are rare and the
hypothesis that invokes the fewest number of these rare events should be
preferred. As long as the rate of evolutionary change is relatively low and can
be assumed to be fairly equal among lineages, parsimony is probably a
reasonable assumption (Felsenstein, 1983).
Other criteria for evaluating trees exist (See Harvey and Pagel, 1991).
Compatibility analysis selects the tree or trees based on the largest possible set
of non-homoplasious characters (Meacham and Estabrook, 1985). Compatibility
analysis has been criticized for ignoring the potential that homoplasious
characters may still carry some phylogenetic signal (i.e. some characters that
could contribute meaningful information to the analysis would be ignored).
Phylogenetic Investigator 10
Maximum likelihood is another method that uses estimates of the probability for
each possible evolutionary event to estimate the tree with the highest
probability of having been produced. Maximum likelihood can be used where the
assumptions required for parsimony are not valid.
Phases of Phylogenetic Inference
Phylogenetic inference can be divided into 4 phases: selection of ingroup
and outgroup, identification of characters and states, assignment of polarity,
and phylogenetic tree construction.
Selection of Ingroup and Outgroup
In scientific practice, the identification of the ingroup, or the group of
taxa to be studied, is usually determined by a systematist who begins with a
particular group of problem taxa in mind. Usually, it is assumed that the larger
taxa are already monophyletic (Eldredge and Cracraft, 1980) and that the goal
of analysis will be to establish the relationships of the ingroup. If these
relationships are uncertain a lower-level study may be undertaken first to
resolve uncertainty about the in-group. Lower-level studies often use large
numbers of taxa to look for groups that appear to be monophyletic (Stevens,
1991)
The definition of the ingroup constrains selection of the outgroup. The
outgroup" consists of taxa selected to determine which states of characters are
ancestral or derived. The most desirable outgroup is the most closely related
taxon to the ingroup, but in the event that this is unknown, any closely related
species that are not within the ingroup can be selected (Stevens, 1991).
Determination of Characters and States
Any set of non-identical taxa can be divided by separating those that
possess any feature "A", and those that do not. Any such feature can be used
as a character for phylogenetic inference. For example, some plants contain
enzyme A and some do not. "Enzyme A" would be the character and "present"
and "absent" would be the two states of the character.
Some features do not seem to have just two states. For example, if we
collected some evergreen branches, we might see that some have bundles of
needles containing 1, 2, or 5 needles. This kind of multistate feature can be
coded as a series of two binary characters in two different ways based on what
is believed about the evolutionary sequence of events. If it is believed that 1 is
ancestral to 2 and 2 is ancestral to 5 (1 -> 2 -> 5), then the first binary
character will derived for those taxa with either 2 or 5 and the second binary
character will be derived only for those with 5. If 1 is considered ancestral for
both 2 and 5 (2 <- 1 -> 5) or if the sequence is unknown, then the first binary
character will be derived only for those taxa with 2 and the second only for
those with 5.
Phylogenetic Investigator 11
Assignment of Polarity
The assignment of character states as ancestral and derived, termed
"polarity," is perhaps the most crucial step of phylogenetic inference.
Phylogenetic methods require groupings based only on derived characters.
Therefore, it is critical to be able to recognize them when they occur.
Characters that have phylogenetic information will only contribute to the
finished hypothesis if they are correctly polarized.
There are several methods for determining the polarity of characters.
Three of the most important methods are outgroup, paleontological, and
ingroup (Stuessy and Crisi, 1984). Each method has its strengths and
weaknesses. Each can explain certain types of data and each has methods for
explaining conflicting data. For all of the methods, conflicting data will be
explained as homoplasy (convergent evolution) during tree construction.
Outgroup method
The outgroup method of determining polarity of character states is
probably the most use commonly used. For each character, the state in which it
exists in the outgroup is considered ancestral and the other state is derived.
This method is based on the generalization that characters that have become
derived for the ingroup will probably not be derived in a closely related group
that diverged prior to the common ancestor of the taxa in the ingroup. The
outgroup method can account for conflicting data by reevaluating whether
some outgroups should be considered part of the ingroup or vice versa. The key
to successful use of the outgroup method is to have well-resolved groups:
knowledge about relationships among taxa in the outgroup improves the ability
to estimate the ancestral state of characters for the ingroup. (See Maddison et
al, 1984 for a more comprehensive description).
Paleontological method
The paleontological method uses fossil taxa for the outgroup. The state
in which each character exists in the outgroup is considered ancestral and the
other state is derived. Although one might think that fossil evidence could
resolve all questions about the polarity of characters, there are two reasons
why it does not: First, it is impossible to determine whether fossils represent
taxa which are direct ancestors of living taxa or a taxon which diverged from
the lineage leading to the present taxa. For this reasons, fossils should be
treated essentially the same way as outgroups. Second, fossils often can not be
accurately coded for many of the characters described from living taxa. Many
features of organisms, like behavior, cannot be easily inferred from fossil
evidence even under ideal conditions and fossils are often fragmentary and
incomplete.
If the fossils are close in temporal position to ancestors of recent species
and if a significant percentage of the characters can be unambiguously coded,
then fossils can greatly improve the resolution of ancestral character states.
Phylogenetic Investigator 12
The paleontological method can account for conflicting data through appeals to
the incompleteness of the fossil record.
In-group method
The in-group method is probably the weakest of the criteria described
here. The most common form of a character among the ingroup is considered
ancestral. For example, if 5 taxa have state A of character 1 and 3 taxa have
state B, then state A is considered ancestral. This method is based on the
generalization that the most common character states among the in-group
represent the primitive condition. Older, larger, and diverse groups are less likely
to preserve the primitive state as the most common character (Stuessy and
Crisi, 1984). The in-group method is most useful as a form of corroboration or
for use when other methods provide ambiguous results.
Tree Construction
Using parsimony, phylogenetic tree construction is a search among
possible arrangements of relationships among taxa and characters that result in
the fewest possible transitions of character states. For any data set, there are a
finite number of possible arrangements of taxa and characters. For data sets
with very few taxa, it is possible to construct all possible trees and see which
require the fewest number of steps (transitions). The number of possible trees
grows exponentially with the addition of taxa, however, and this method quickly
becomes impractical to perform by hand. There are, however, strategies and
heuristics which can allow the problem-solver to greatly limit the number of
possibilities which must be considered. In most problems, only a few trees are
actually supported by any of the data.
Each character in the data set, defines a group of taxa potentially
descended from a postulated ancestor, and therefore can be seen as direct
support for the existence of a postulated common ancestor or node. The real
set of possible trees, then consists only of those trees which could be
constructed from the available nodes.
Characters are inclusive/exclusive when they define identical, nested, or
exclusive groups. For example, assume that character 1 defines a group of
{R81, R82, and R83}. If another defines the same set of taxa, the characters
are identical characters. If another character defines a subset or a superset of
characters (e.g. {R81 and R82} or {R81, R82, R83, and R84}), the characters
are nested with respect to each other. If another character defines completely
different set of taxa (e.g. {R85 and R86}) the characters are exclusive with
respect to one another. Characters conflict when they overlap incompletely. For
example, assume that character 1 defines a group of {R81, R82, and R83} and
character 4 defines a group of {R82, R83 and R84}. These two groups are
contradictory because each character claims some, but not all of the taxa of
the other. Character compatibility groups can be formed that place some or all
of the characters into a hierarchical arrangement to evaluate how many of the
Phylogenetic Investigator 13
characters will support a particular hypothesis (arrangement of the taxa) and
how many extra steps will be needed to account for incompatible characters.
Ideally, all of the characters will agree in defining a single tree. In practice,
some characters will define contradictory groups (groups that overlap
incompletely). The largest possible group of inclusive/exclusive characters can
serve as a working hypothesis from which to construct a phylogenetic tree. This
tree can then be optimized for parsimony if so desired.
A phylogenetic tree is a branching path from a single point at which all of
the character states are ancestral to several points where they are the same as
the taxa in the ingroup. The lowest node, the node at the bottom of the tree,
will be entirely ancestral, The postulated node above that will be linked to the
lower node and will have a transition or transitions. Its states, then, are partially
ancestral and partially derived. If it has the same states as any of the ingroup,
they can be directly linked. The next postulated node has more derived states
and may be linked to more recent taxa, until all of the taxa have been
accounted for.
Constructing the phylogenetic tree involves adding postulated ancestors
for each of the unique inclusive/exclusive characters, linking the ancestors
together and to the taxa in the ingroup, adding the transitions for the
characters which support the structure, and then distributing the homoplasious
(conflicting) characters either as parallel gains or gains with subsequent
reversals. (I suggest initially adding homoplasious characters as parallel gains,
wherever possible. This makes it easy to spot duplicated characters each of
which should be considered in order to evaluate alternate topologies and
character optimizations.)
Once a tree has been constructed, it can be assessed and, if necessary,
revised to ensure that it is a minimum length (most parsimonious) tree. Tree
assessment should begin by examining each homoplasious character, beginning
with the one that requires the most transitions, and considering (1) how many
steps could be saved by "fixing" the character (rearranging the tree so that this
character would have a single transition) and (2) how many more steps would
be required in each other character that would be affected by those changes. If
an arrangement is found that results in fewer steps, the tree should be
restructured and then assessed again from the beginning. If an arrangement is
discovered that results in an equal number of steps, assessment should
continue until it is confirmed that no better tree is possible, and then all equally
parsimoniously trees should be reported. The most difficult part of phylogenetic
inference is assuring that all most parsimonious trees have been discovered.
Rigorous assessment and systematic consideration of each homoplasious
character provides the best probability of success.
For each most parsimonious tree, there should also be consideration of
alternate character optimizations. Each homoplasious character should be
considered for how it could be distributed on each most parsimonious tree. One
Phylogenetic Investigator 14
of the most important aspects of the interpretation of phylogenetic trees
involves describing alternate hypotheses that could explain the data set and
suggesting subsequent investigation that could provide insight into these
uncertainties.
Phylogenetic Investigator 15
AN EXAMPLE PROBLEM USING PI
This example problem deals with a set of imaginary insect taxa among
which several wing characteristics vary. Using diagrams of their wings as a data
source, this guide will illustrate how to use PI to determine characters and
states, assign polarity, and construct the most-parsimonious phylogenetic
trees. This example is constructed to allow the reader to follow along using
Phylogenetic Investigator by following the instructions given in italics. Program
structures like windows, menus, and commands are printed in boldface.
For this example, I have selected only a subset of taxa (Fig. 5) from the
data source (see Appendix B for the complete set of taxa). Taxa R04, R08,
R11, R12, and R15 will be the ingroup. We will use R10 as an
R04
R12
R08
R15
R10
F95
R11
Figure 5.
A set of taxa presented as an example problem of
phylogenetic inference. R04, R08, R11, R12 and R15 are
the ingroup, R10 and F95 are used to determine polarity
by the outgroup and paleontological methods.
outgroup and F95 as a representative fossil. The decisions to use these
particular taxa have been made more or less arbitrarily, in order to illustrate
certain aspects of problem solving using PI. Ideally the ingroup will be composed
of all of the taxa descended from some postulated ancestor and the outgroup
will be the sister taxon, or the most closely related taxon not within the
ingroup. In practice, one is constrained by current knowledge and the availability
of study material. Our problem, then, is to define the system of evolutionary
relationships among the ingroup. Having defined our problem, we are ready to
start PI.
Phylogenetic Investigator 16
Double click on the program icon and, after the program finishes opening,
select Set-up Problem from the Problems menu. This causes the Set-up
Problem window to open which contains a scrolling list of taxa (Fig. 6).
Figure 6.
The Set-up Problem window. This window is opened by
using the Set-up Problems item in the Problems menu.
Hold down Command key and select R04, R08, R10, R11, R12, R15, and
F95. Click Add and then click Done. (Note that one could also select a single
taxon, click Add, and repeat until all the desired taxa have been selected and
then click Done). The recent and fossil nodes should appear in the drawing field
and a new window, entitled Chars & States should open directly over them
(Fig. 7).
At this point, we are ready to start identifying characters and states. We
notice that some wings have spots and some don't. At this point, we need not
be concerned which state is ancestral and which is derived. Simply enter the
character and the two states. Click in the top field of the Chars & States
window. Type "Spots" into top field and press tab -- this makes the Ancestral
field active. Type "present" into active field and press tab -- this makes the
Derived field active. Type "absent" into active field and press tab -- this moves
the insertion point back up to the top field.
Click the zoom button at upper right hand corner of Chars & States
window. This transforms the Chars & States window into a spreadsheet type
format. The Chars & States window can be toggled between these two
modes at any time and either window can be used for entering, modifying and
deleting characters.
Phylogenetic Investigator 17
Figure 7.
The compact version of the Chars & States window. Enter characters into
the top field. Once entered they appear in the scrolling list. Enter states
into the lower fields. The three buttons at the bottom allow exchanging
character state names (left button), reversing polarity of data in the data
matrix (right button), or both (middle button).
We notice that some wings have a little branch at the end of the veins
and some don't. Click in the left most field of line 2 (the character field). Type
"Vein branching" and press tab -- this moves the insertion point to the
Ancestral field. You may notice that PI replaces any spaces within characters
and states with underline characters. Type "present" and press tab -- this
moves the insertion point to the Derived field. Type "Absent" and press tab -this moves the insertion point to the next character field.
Enter the rest of the data as it appears in the Table 1. After entering all
the data, click the zoom button at the upper right hand corner of the window.
This will transform the Chars & States window back to the compact
configuration in preparation for assigning polarity.
Phylogenetic Investigator 18
Table 1. Six characters and unpolarized states for the insect wing example.
Once all of the data has been entered, we're ready to start assigning polarity to
the character states. Select the first line in the scrolling field in the middle of
the small Chars & States window. This will bring up the two states assigned
to it in the lower fields (Fig. 8).
Figure 8.
When a character is selected from the scrolling list, the
states for that character can be modified or polarized. In
this figure character 1 has already been polarized.
At the bottom of the Chars & States window are three buttons. The
button on the left exchanges the words for the states in the Chars & States
window. The button on the right inverts the coded data in the data matrix for a
Phylogenetic Investigator 19
character (exchanges 1's and 0's for a whole column). The button in the middle,
labelled Invert Polarity, does both.
By looking at our data source, we see that spots are present neither in
the outgroup (R10) nor in our fossil taxon (F95). Therefore, we will reverse the
polarity of this character.
Press the left-hand button. This will exchange the two character state
words -- after pressing the button your window should match Figure 7. As we
look at the rest of the taxa we can see that some are already polarized
correctly and others need to be exchanged.
When we get to character 4, we realize there is a problem. Character 4 is
present in the fossil, but absent in the outgroup. In this case, we can use the
ingroup method to evaluate which should be ancestral: it is present in 2
members of the ingroup, but absent in the other 3, therefore absent should be
considered ancestral.
Polarize the rest of the characters. When you have polarized all your
characters, they should match the table below.
Table 2. The characters and polarized states for the insect wing example.
Having finished polarity, we are ready to open the Data Matrix and code
the data (Fig. 9). Select the Data Matrix item from the Windows menu. When
the Data Matrix is initially opened, there should be a row for each taxon and a
column for each character. These should all be 0's, unless the right hand Invert
Polarity buttons have been used.
In the Chars & States window select character 1. Look at each taxon in
turn, determine whether or not it possesses the ancestral or the derived
condition for the character. If the taxon has the derived condition, click on the
symbol where the row for that taxon and the column for character 1 intersect.
This will cause the symbol to change from the ancestral "0" to the derived "1".
A second click will cause it to toggle back. Code the rest of the data by
selecting each character in turn and considering each taxon. At this point, we
Phylogenetic Investigator 20
are finished with and can close the Chars & States window. by selecting the
Chars & States item from the Windows menu.
At this point we begin phylogenetic tree construction and begin to search
for any patterns in the data matrix that indicate phylogenetic signal. In order to
increase our ability to recognize patterns, we can organize the taxa more
effectively and as we find patterns that appear to indicate phylogenetic signal,
we can also restructure the matrix to aid recognition and memory. Organizing
the taxa in the matrix as described here is not necessary for tree construction,
but it can greatly aid finding patterns among the data.
Although taxa can be moved up and down in the data matrix at any time,
characters can only be moved when no links are selected. Click on taxa to move
them. This brings up a horizontal box which highlights the row to be moved and
changes the cursor to a sideways arrow. Click between the two lines where the
taxon is to be moved. To move a
Figure 9.
The Data Matrix window. Each row in the matrix
represents the data for a taxon and each column represents
a character. Characters are coded with symbols for
ancestral (0) and derived (1) states.
character, click on the column heading when no link is selected and a vertical
box which hilites the column is displayed. Click on a second column heading and
the character is moved into that column.
Phylogenetic Investigator 21
Figure 10.
The initial arrangement of the data matrix.
After initial inspecting the original data matrix (Fig. 10), we can make a
change that will enhance our ability to recognize patterns: we can move the
outgroup (R10) to the bottom of the matrix. This will separate the ingroup and
outgroup taxa. Click on R10 and then, with the sideways arrow cursor, click
between the rows where you want the taxon to appear -- in this case, just
above F95 (Fig. 11).
Figure 11.
R10 has been moved together with F95 separating the
ingroup and outgroup taxa.
Now we can exclusively consider relationships within the ingroup. First, we
notice that 6 and 3 have the opposite pattern. These characters are
"inclusive/exclusive". If we put the 1's in character 6 together, we may be able
to emphasize this pattern. Bring R15 up to just below R04 to put the 1's in
character 6 together (Fig. 12).
Figure 12.
R15 has been joined with R04 on the basis of character 6.
We can see now that 3 and 6 are exclusive from each other and can both
be nested within 2. We can move 6 to the other side of 2 so as to emphasize
Phylogenetic Investigator 22
that pattern (Fig. 13). Click on the column heading for character 6. Once it is
outlined, click on the column heading for character 2.
Figure 13.
Character 6 has been moved to the other side of character 2
from character 3 to emphasize this division of the taxa.
Now we can see that 4 (disregarding the outgroup problems) and 5 nest
nicely within 3. We can also see that 1 just doesn't fit at all. 1 conflicts with 6
and 4 and 5. Move 1 to outside the group of inclusive/exclusive characters to
set it apart (Fig. 14).
Figure 14.
Character 1 has been separated from the other taxa to
separate homoplasious and non-homoplasious characters .
The organization of this matrix now represents an inclusion/exclusion
hypothesis. It shows us that R04 and R15, based on sharing character 6, will be
a group separate from R08, R11 and R12 (which share character 3). Also, we
can see that the group of 3 taxa will contain a subgroup composed of R11 and
R12 (because they share 4 (with homoplasy in F95) and 5. Now, with our
completed inclusion/exclusion hypothesis, we're ready to draw some
phylogenetic trees.
Phylogenetic Investigator 23
Figure 15.
To make a link, select a second node while pressing the
shift key. (Or press the shift key and select two nodes).
First, we can move the recent taxa at the top to represent the order
described in the Data Matrix. F95 will at the extreme left and R10 at the left of
the recent taxa. Then R04 and R15 will be together, then R08, and then R11
and R12. Within the two subgroups R04, R15 and R11, R12, order is not
significant. This order will produce a diagram which appears to have a trend of
increasing numbers of derived characters from left to right. This trend is
actually an illusion: the branches could be arranged such that R04, R15 was on
the right of R08, R11, R12. Nevertheless, it is often useful to use a consistent
form of representation because it can facilitate both construction and
interpretation.
Select Add Node from the Actions menu, and click near the bottom of
the screen. This node will be our outgroup node. When the node appears it is
selected. Because the outgroup node and the outgroup have the same
distribution of characters states (all ancestral), they can be immediately linked.
Holding the shift key down, we click on R10 (Fig. 15). This forms a link and
unselects both nodes. Note that R10 is connected to PA with a vertical line.
This indicates qualitatively, in addition to the fact that no transitions will appear
on this line, that there are few or no differences between the ancestor and this
descendant taxon.
Phylogenetic Investigator 24
Figure 16.
Click on a link to select it. Press the column heading in the
data matrix while a link is selected to add a transition to a
link.
We then create a second node. This node will be the ingroup node, from
which all the taxa in the ingroup (all the taxa that share character 2) are
descended. After linking this node to PA, click on the link, selecting it, and then
click on the character 2 column heading in the Data Matrix. This will add a
forward transition for character 2 to the selected link (Fig. 16). Note that there
are no taxa which possess only character 2, so PB should not be linked directly
to any taxa in the ingroup.
Phylogenetic Investigator 25
Time
0
R10
R04
R15
R08
R12
R11
1 O>l
1 O>l
10
1 O>l
PE
PC
4 O>l
5 O>l
20
6 O>l
PD
3 O>l
30
F95
PB
40
4 O>l
2 O>l
50
PA
Morphological Change
Figure 17.
In this most parsimonious tree, character 1 is distributed
as 3 convergent forward transitions (in R04, R08, and
R12).
We can then add a node (PC) under R04, R15 for character 6, a node
(PD) under R08, R11, R12 for character 3 and a node (PE) under R11, R12 for
characters 4 and 5. We can then link up all the taxa (eventually linking F95 also
to the outgroup node with a homoplasious gain for character 4). We are then
left with character 1. Character 1 can be added as 3 separate gains in R08,
R12, and R15. This implies that character 1 evolved separately three times (Fig.
17).
This optimization of character 1 provides an avenue of subsequent
research. If character 1 evolved three separate times in recent history, perhaps
some major climatic or environmental change occurred where these taxa occur.
Perhaps a new predator appeared or arrived. Perhaps these taxa invaded new
areas that placed similar constraints on evolutionary development. This
optmization of character 1 predicts that if we discover fossil taxa closely
related to PB, PC, PD and PE, none of them will have character 1 in the derived
state. All of these are avenues for gaining further insight into character 1.
Phylogenetic Investigator 26
Time
0
R10
R04
R15
R08
R12
R11
1 l>O
1 O>l
10
PE
PC
4 O>l
5 O>l
20
6 O>l
PD
1 O>l
3 O>l
30
F95
PB
40
4 O>l
2 O>l
50
PA
Morphological Change
Figure 18.
In this most parsimonious tree, character 1 is distributed
as 2 convergent forward transitions (in PD and R15) and a
reversal (in R11).
Character 1 can also be two gains (in R15 and PD) and a loss (in R11)
(Fig 18). To generate this optimization from the previous arrangement, select
link R12-PE and click the character 1 button twice. This causes the transition to
change first to a reversal and then to be removed entirely. Do the same for link
R08-PD. Then select link PB-PD and click (the character 1 button) once -- this
adds the forward transition. -- and select link R11-PE and click (the character 1
button) twice. This adds a reversal for character 1.
This optmization of character 1 predicts that if we discover fossil taxa
closely related to PD, it will have character 1 in the derived state, but that taxa
closely related PB and PC will not. Biogeography might again offer insights into
parallels between R15 and the other taxa.
Phylogenetic Investigator 27
Time
0
R10
R04
R15
R08
R12
R11
1 l>O
1 l>0
10
PE
PC
4 O>l
5 O>l
20
6 O>l
PD
3 O>l
30
F95
PB
40
4 O>l
1 O>l
2 O>l
50
PA
Morphological Change
Figure 19.
In this most parsimonious tree, character 1 is distributed
as 1 convergent forward transitions (in PB) and two
reversals (in R04 and R11).
Character 1 can also be 1 gain (in PB) and two losses (in R04 and R11)
(Fig. 19). To generate this optimization from the previous arrangement, select
link PC-PD and click the character 1 button twice. This causes the transition to
change first to a reversal and then to be removed entirely. Do the same for link
R15-PC. Then select link PA-PB and click once -- this adds the forward
transition. -- and select link R04 PC and click twice. This adds a reversal for
character 1.
This optmization of character 1 now focuses attention on the taxa which
appear to have lost character 1. Is there some environmental or biogeographical
factor that can be associated with the loss? Now, if we discover fossil taxa they
should all have character 1 in the derived state.
Phylogenetic Investigator 28
Time
0
R10
R04
R15
R08
R12
R11
1 l>O
10
4 O>l PE
5 O>l
20
3 O>l
6 l>O
PD
30
F95
40
1 O>l
PC
PB
4 O>l
2 O>l
6 O>l
50
PA
Morphological Change
Figure 20.
In this most parsimonious tree, character 1 is distributed
as 1 forward transition (in PF) and 1 reversal (in R11).
Saving a step in character 1 is achieved by explaining
character 6 using 2 steps -- a forward transition (in PB)
and a reversal (in PD).
It is also possible to construct a second topology which improves
character 1 by a step, but adds a step to Character 6 (Fig. 20). Character 6 is
then gained in PB and lost in PD and Character 1 is gained in PF and lost in R11.
To construct this topology, select link PB-PD and select Reassign Link from
the Actions menu. Use the pop-up menu PB to change the node assignment to
PC. Then select link R04-PC and Reassign Link from PC to PB. Instead of using
the menu command, it is also possible to select the link and hold down the shift
key while selecting the node to be reassigned. This causes the pop-up menu to
appear right on the drawing field.
Having constructed a phylogenetic tree or a series of phylogenetic trees,
interpretation is necessary for them to become meaningful. Each speciation
event and each character transition should be considered thoughtfully from a
historical perspective: What was the environment? What other evidence
(ecology, biogeography, etc.) might support or contradict the evidence used to
construct the tree? The homoplasious characters are of particular interest: are
these characters highly variable among other taxa? Is it possible to look at the
Phylogenetic Investigator 29
character more closely to investigate how it has been defined? Does the
homoplasious character vary in function across groups?
If we were dealing with plants, rather than insects, we might be asking
whether some of the character incompatibility observed was due to the
presence of hybrids. Hybridization is rare among animals, but often causes
problems for phylogenetic inference with plants because hybrids may share
characteristics of taxa from different lineages. Alternatively, derived characters
are often recessive and some hybrids may have no derived characters at all.
Hybrids can be dealt with in a variety of ways. One way is to simply remove
them from the sample. Hybrids are not really taxa in that they often cannot
themselves reproduce. Another way is to place them with links between them
and the taxa from which they are derived.
Phylogenetic Investigator 30
PHYLOGENETIC INVESTIGATOR REFERENCE MANUAL
Phylogenetic Investigator (PI) is designed to facilitate modeling and
practicing fundamental phylogenetic inference. We believe that beginning
students of phylogenetic inference should be able to (1) inspect the data, make
inferences, and build representations one step at a time, (2) vary
representational features of their trees (such as angle of divergence and time
between speciation events), (3) create reticulate tree patterns, and (4) view all
of the character transformations at one time. No other available software
package allows students to do any of these things. It was for these purposes
that we created Phylogenetic Investigator.
PI provides tools for managing and manipulating up to 20 characters of
binary phylogenetic data for 15 or fewer taxa. PI has been designed with 2 data
sets in mind: the Caminalcules and the Dendrogrammaceae, but other data sets
can be adapted for use (See the section on Set-up Problems below). With PI,
students can wrestle with the assumptions, methods, goals, and limits of
phylogenetic inference. Once students have become conversant with the
concepts and functional relationships implied by phylogenetic inference other
more research-oriented tools may be better suited. More advanced tools can
allow students to use more complex transformation series, weight characters,
and experiment with the effects of including and excluding characters and taxa.
The guide to PI below is organized systematically to facilitate finding
information about particular features of the program. Windows are described
first and then menus. Dialog boxes are described with the menu item that opens
them.
Windows
PI uses two windows for data management (Chars & States and Data
Matrix) and one for tree construction (Phylogenetic Tree). Most will open
automatically when a problem is selected or set-up. None of these windows
have close boxes and must be opened or closed using the Windows menu.
Chars & States
The Chars & States window will open automatically if Set-up Problem has
been used to pose a problem. This window has two configurations and the user
can move between them by clicking the zoom button at the upper right hand
side of the window. Data can be entered using either configuration and the
small configuration (Fig. 21) can be used for polarizing characters.
Phylogenetic Investigator 31
Figure 21.
The compact version of the Chars & States window. Enter characters into
the top field. Once entered they appear in the scrolling list. Enter states
into the lower fields. The three buttons at the bottom allow exchanging
character state names (left button), reversing polarity of data in the data
matrix (right button), or both (middle button).
Small configuration
Upon opening, the upper left field should be active. The user enters Characters
here, causing them to be entered into the list of characters below. As
characters are entered here, a column is automatically created in the Data
Matrix window for coding. A total of 20 characters can be defined. The user is
automatically prompted to enter first the ancestral and then the derived state.
All of these fields can contain only a single word and the program will
automatically substitute underline characters for spaces, if entered.
Items in the list can be modified by shift clicking -- this will bring up a
dialog that asks what the new item should be. Items can also be deleted by
option clicking -- this will bring up a warning/confirmation dialog. By selecting
different characters from the list, one can subsequently modify states for that
character.
At the bottom of this window are three buttons. The middle button,
labelled Invert Polarity, exchanges the terms entered for ancestral and derived
characters and also exchanges 1's and 0's in the column for that character in
the Data Matrix. The button to the left only exchanges ancestral and derived
terms and the button on the right only inverts the polarity of the column in the
data matrix.
Phylogenetic Investigator 32
Large configuration
In this mode, the window has a spreadsheet type format (not pictured).
Characters and states can be entered, but only in order. A tab will move the
insertion point to the next active field. A return will move the insertion point
down one row (if that row is active). If a character is deleted, the user is asked
to confirm deletion before the line of data from the data matrix is removed.
Data Matrix
The Data Matrix (Fig. 22) is a palette, meaning that this window will float
over all the others. It is often useful to move this window to the right so that
only that portion which contains data is visible. There are three fields in this
window. The Problem field at the bottom shows the title of the problem that is
currently being addressed. This field will be filled in automatically when a model
or practice problem has been selected, but it is also user modifiable. The
contents of this field is what is used as the default file name when a problem is
saved for the first time. This field also communicates with the problem field in
the expanded Chars & States window. The small field in the upper left shows the
current tree length (in unweighted transitions). The large, central field contains
the data matrix currently being used for problem-solving.
Figure 22.
The Data Matrix. Data consists of 1's for ancestral and 0's for derived
character states and is organized with taxa in rows and characters in
columns. At the upper left is the number of unweighted transitions in the
tree. The field at the bottom is user modifiable and contains the name of
the problem.
In the data matrix, characters are in columns and taxa are in rows. When a
link between nodes is selected in the tree construction window, a click on a
Phylogenetic Investigator 33
character button (in the row above the matrix) will add a transition for that
character to the selected line. A second click will change the transition into a
reversal and a third click will remove the
transition from the line. The tree length field is updated automatically. States
for taxa can be modified by clicking on the state character for a taxon. This will
toggle between the ancestral and derived characters. Holding down the option
key and clicking allows one to change the character to X to indicate missing
data. Rows can be moved by clicking on them, which will bring up a box
outlining the row to be moved and different cursor. A second click, indicating
where the row should be moved to (between rows or above or below another
row) will move the row to this location. Columns can be moved by clicking on a
character number above the data table while no line is selected on the
phylogenetic tree. This will reveal a box outlining the column to be moved. Click
on another character button to move the column into that space in the matrix.
Phylogenetic Tree
In Phylogenetic Investigator, trees are constructed from nodes, links, and
transitions. Nodes and links can be selected by clicking on them. To de-select
everything, click on the background. Nodes can be moved by dragging. To form
a link, use the shift key to select two nodes. These nodes will be automatically
linked and the link will subsequently follow the nodes if moved. Transitions are
added to links by clicking on the character buttons in the Data Matrix window.
About Nodes
All organism designations (Nodes) begin with letters that indicate the
organism's status R for recent, F for Fossil, and P for Postulated. Recent and
Fossil organisms are numbered and can be constrained temporally (this property
is controlled by the Time checkbox in the settings window). Postulated
organisms have sequential letters are free to move in both axes. When nodes
are selected, they can be deleted by using the Remove Nodes menu item. All
associated links will also be removed (this is sometimes a fast way to
reconstruct a tree for a revision). Holding the shift key down allows two nodes
to be selected. Once a second node has been selected, a link is formed between
them and both are de-selected. Holding the shift key down and selecting a node
while a link is selected brings up a pop-up menu that allows reassigning the link
from the selected node to any other node. If a node is selected and the Add
Node command is executed while holding the shift key down, a new node will be
added and linked to the previously selected node.
Phylogenetic Investigator 34
About Links
Links can be selected by clicking on them. Selected links can be removed
or reassigned (by using menu items). Selected links can have transitions
assigned to them by clicking on the character button in the Data Matrix. Holding
the shift key down and selecting a node while a link is selected brings up a popup menu that allows reassigning the link from the selected node to any other
node.
Settings
The settings window (Fig. 23) allows the user to modify the time scale on
the phylogenetic tree, change the characters used for ancestral and derived
characters, and to apply or remove temporal constraint from a problem. The
temporal constraint is turned on by default. If turned off, it will remain off until
turned on again (even between uses of the program).
Figure 23.
The Settings window. The time scale and constraint may be
modified during problem-solving. Modifying the Matrix
Symbols during problem-solving may result in erratic
behavior.
The matrix symbols currently in use are the uppercase letter 'O' (as in
Oliver) and lowercase letter 'l' (as in lollipop). These were what I thought looked
the best after trying many other possibilities. (Real 1's and 0's don't line up
right vertically as nicely as O's and l's.)
Note: Changing matrix symbols during problem-solving is probably a bad
idea. It might not be fatal, but could cause some odd behavior with transitions.
Menus
Apple
The Apple Menu contains the About Phylogenetic Investigator item which
opens the Phylogenetic Investigator splash screen.
Phylogenetic Investigator 35
File
New
This clears the drawing field, data matrix, and characters and states.
Open...
This item will open a PI Treefile
Save
Save As...
These items generate a PI Treefile. Treefiles contain a snapshot of the
current state of the problem: Characters, states, coded data, nodes, links,
locations, and transitions.
Open Nexus
This feature has not yet been implemented. Look for it in future versions
of PI.
Save Nexus
This saves the current data in a form which can be read by PAUP and
MacClade 3.x.
Save MacClade 2.1
This saves the current data in a form which can be read by the older
version of MacClade.
Export Tree
This item creates a ClarisWorks PICT file with the current tree and Data
Matrix.
Print...
This opens a dialog box (Fig. 24) with two radio buttons and three
checkboxes. One can select to print the data as a practice problem or as
a setup problem. As a practice problem, the data matrix and phylogenetic
tree are put together on a single page and printed. As a setup problem,
one can select phylogenetic tree, data matrix, and characters and states
for printing. Each will appear on a separate page.
Figure 24.
The printing dialog box.
The phylogenetic tree and data matrix are printed exactly as they appear
on the screen. The Characters and States are automatically transferred to
a form for printing.
Phylogenetic Investigator 36
Quit
This item retains the current problem and quits the application
Edit
Cut, Copy, Paste, and Clear are implemented.
Actions
Add Node
When this item is selected, the cursor changes to appear like a postulated
node and when the mouse is clicked, a new postulated node is placed at
that point and selected.
Remove Link
If a line is selected, this command will remove it and updates tree length
if transformations were present on the link removed. Links can also be
deleted by pressing the delete key.
Remove Node
If a node is selected, the program confirms and then removes the
selected node and attached links. Nodes can also be removed by pressing
the delete key.
Reassign Link...
If a line is selected, this command will open a dialog box (Fig. 25) with a
line and two pop-up menus. Select the pop-up menu for the end of the
line to be moved and select the node it is to be reassigned to. Selecting
either of the nodes that already terminate the line, or clicking the cancel
button, will cancel this command and close the window.
Figure 25.
The Reassign Link dialog box.
Links can also be reassigned by selecting a line, holding down the shift
key and selecting one of the nodes at either end of the line. This will
cause a pop-up menu to appear at that node. Selecting one of the nodes
Phylogenetic Investigator 37
from the menu will cause that end of the link to be reassigned to the
selected node.
Problems
There are three types of problems that can be selected under the
problem menu. At the top of the menu is the Set-Up Problem...
command which opens a dialog box and allows the user to define a set of
organisms for a problem. The lower two sections of this menu provide
tree construction problems for students which are useful for learning the
mechanics of tree construction prior to addressing determining characters
and states and assigning polarity. The second area of the menu contains
Model problems. These problems always display particular characteristics,
but the specific taxa and the arrangement of the characters will vary each
time. The lowest area on the menu contains 5 problems of generally
increasing complexity. Each of these problems will display similar
characteristics each time it is selected, but may produce substantially
different results.
Set-Up Problem
Opens the Set-up Problem dialog box (Fig. 26). Select the taxa from the
scrolling list and Add them to the problem set. When complete, select
Done and the selected taxa will be placed in the tree
Figure 26.
The Set-up Problem dialog box.
construction window. Non-contiguous selections can be made by using
the Command (cloverleaf) key.
The taxa listed here represent the Caminalcules (R1-29, F1-77) the
Dendrogrammaceae (R1-18), and the model problem taxa (R80-R89, F90Phylogenetic Investigator 38
F99). Other sets of taxa can be adapted for use within PI by assigning a
label for each one. For recent taxa either R1-29 or R80-89 can be used.
For fossil taxa less than 50 million years old, F90-F99 can be adapted
(they appear in pairs at 10 million year intervals). Future versions of PI
may permit modification of the taxon data base.
Each taxon that is added here will be given a line in the data matrix
for coding character and state data. The software can accommodate up
to 15 taxa in a problem set. It is not recommended to construct problems
with more than this number of taxa.
Taxa can be added at any time during the problem solving process.
Taxa added after characters have been defined will be coded with an "X"
for each character. Note: It is nonsensical and not-advised to add taxa to
a model or practice problem.
Model Problems
The second area defined in the Problems menu contains a list of
predefined problems: Autapomorphy; Synapomorphy 1, 2, and 3; and
Homoplasy 1 & 2, 3, and 4. Each of these problems, when selected, will
produce a data matrix and add several taxa to the drawing field. In every
case, the taxa selected and the order of the taxa and characters in the
matrix will be randomized, but the form of the resultant phylogenetic tree
will be the same each time. An example problem with solution and
comments is provided for each model problem in Appendix B.
Practice Problems
Like the model problems, the practice problems randomly select and
arrange a group of taxa and characters each time they are selected.
These problems show much greater variability than the model problem.
Problem 5 has two parts. After solving the first part, select Problem 5b
and an additional taxon with data is added to the problem.
Windows
Each menu item simply opens the window named (or brings it to the
front, if hidden or closed).
Phylogenetic Investigator 39
OTHER SOFTWARE FOR PHYLOGENETIC ANALYSIS
There are many sets of software tools for phylogenetic research. Three of
the most important are MacClade, PHYlogenetic Inference Package (PHYLIP),
and Phylogenetic Analysis Using Parsimony (PAUP). Most packages now allow
some form of automated searching for trees that meet various criteria (tree
length, etc.).
MacClade (Maddison & Maddison, 1989) is a well designed Macintosh
software package which allows the user to evaluate the effects of swapping
branches on the tree length. This is particularly useful for evaluating a series of
closely related hypotheses. An early version of MacClade (2.1) appears on the
BioQUEST CD-ROM and is freely distributable. More recent versions are available
for purchase. All distribution is by Sinauer Associates, Sunderland,
Massachusetts 01375, USA. Their phone number is: (413) 665 3722, FAX:
(413) 665 7292.
PHYLIP (Felsenstein, 1993) is a large set of free programs which appear
to have designed for the UNIX environment, but which have been ported to
Macintosh and DOS platforms. PHYLIP's interface is not very Macintosh-like (for
lack of a better term). PHYLIP is available by "anonymous ftp" over electronic
networks (including the PCDOS, 386 PCDOS, 386 Windows, and Macintosh
executables) from evolution.genetics.washington.edu (128.95.12.41). Contact
Joe Felsenstein <[email protected]> for details or start by fetching
file pub/phylip/Read.Me.
PAUP (Swofford, 1991) is probably the single most widely used package
by researchers. It provides a fairly Macintosh-like interface and allows a wide
variety of options for searching for phylogenetic trees. Previous versions have
been available from the Center for Biodiversity, Illinois Natural History Survey,
607 East Peabody Drive, Champaign, Illinois 61820, U.S.A.
Phylogenetic Investigator 40
SUGGESTED READINGS
For a highly readable treatment of the evolutionary issues relevant to cladistics,
read Stephen Jay Gould's (1989) Wonderful Life: The Burgess Shale and the
nature of history.
For a general account of cladistics, try Mark Ridley's (1986) Evolution and
Classification: The reformation of cladism.
For a thorough and readable introduction to cladistic applications, read Daniel
Brooks and Deborah McClennan's (1991) Phylogeny, Ecology, and Behavior.
For an in-depth treatment of the scientific revolution in cladistics try David
Hull's (1988) Science as a Process.
For a thorough background on phylogenetic diagrams, try Niles Eldredge and
Joel Cracraft's (1980) Phylogenetic Patterns and the Evolutionary Process:
Method and theory in comparative biology.
For a thorough treatment on the philosophy of phylogenetic inference try Elliott
Sober's (1988) Reconstructing the past: Parsimony, evolution and inference..
The English version of the book that started it all is Wili Hennig's (1966)
Phylogenetic Systematics.
Phylogenetic Investigator 41
BIBLIOGRAPHY
Brooks, D. R., & McClennan, D. A. (1991). Phylogeny, Ecology, and Behavior.
Chicago: University of Chicago Press.
Brooks, D. R., McLennan, D. A., Carpenter, J. M., Weller, S. G., & Coddington, J.
A. (1995). Systematics, ecology and behavior. Bioscience, 4 5(10), 687695.
Davis, G. M. (1995). Systematics and public health. Bioscience, 4 5(10), 705714.
de Queiroz, K. (1985). The ontogenetic method for determining character
polarity and its relevance to phylogenetic systematics. Systematic
Zoology, 3 4(3), 280-299.
Duncan, T., Phillips, R. B., & W.H. Wagner, J. (1980). A comparison of branching
diagrams derived by various phenetic and cladistic methods. Systematic
Botany, 5(3), 264-293.
Eldredge, N., & Cracraft, J. (1980). Phylogenetic Patterns and the Evolutionary
Process: Method and theory in comparative biology. New York: Columbia
University Press.
Felsenstein, J. (1983) Parsimony in systematics: biological and statistical
issues. Annual Review of Ecology and Systematics, 1 4, 313-333.
Felsenstein, J. (1993). PHYLIP: Phylogeny inference package. Distributed by the
author. University of Washington.
Gould, S. J. (1980). The Panda's Thumb. New York: W. W. Norton & Company.
Gould, S. J. (1989). Wonderful Life: The Burgess Shale and the nature of
history. New York: W.W. Norton and Company.
Harvey, P. H., & Pagel, M. D. (1991). The comparative method in evolutionary
biology. New York: Oxford University Press.
Hennig, W. (1966). Phylogenetic Systematics. Chicago: University of Illinois
Press.
Hull, D. L. (1988). Science as a Process. Chicago: University of Chicago Press.
Lauder, G. V., Huey, R. B., Monson, R. K., & Jensen, R. J. (1995). Systematics
and the study of organismal form and function. Bioscience, 4 5(10), 696704.
Maddison, W., & Maddison, D. (1989). MacClade: Software for cladistic analysis.
Maddison, W. P., Donoghue, M. J., & Maddison, D. R. (1984). Outgroup analysis
and parsimony. Systematic Zoology, 3 3(1), 83-103.
Meacham, C. A., & Estabrook, G. F. (1985). Compatibility methods in
systematics. Annual Review of Ecological Systematics, 1 6, 431-446.
Miller, D. R., & Rossman, A. Y. (1995). Systematics, biodiversity, and agriculture.
Bioscience, 4 5(10), 680-686.
Phylogenetic Investigator 42
Ridley, M. (1986). Evolution and Classification: The reformation of cladism. New
York: Longman Group Limited.
Savage, J. M. (1995). Systematics and the biodiversity crisis. Bioscience,
4 5(10), 673-679.
Simpson, B. B., & Cracraft, J. (1995). Systematics: The science of biodiversity.
Bioscience, 4 5(10), 670-672.
Sober, E. (1988). Reconstructing the past: Parsimony, evolution and inference.
Cambridge: MIT Press.
Sokal, R. R. (1983a). A phylogenetic analysis of the Caminalcules: I. The data
base. Systematic Zoology, 3 2(2), 159-184.
Sokal, R. R. (1983b). A phylogenetic analysis of the Caminalcules: II. Estimating
the true cladogram. Systematic Zoology, 3 2(2), 185-201.
Stevens, P. F. (1991). Character states, morphological variation, and
phylogenetic analysis: A review. Systematic Botany, 1 6, 553-583.
Stuessy, T. F., & Crisi, J. V. (1984). Problems in the determination of
evolutionary directionality of character-state change for phylogenetic
reconstruction. In T. Duncan & T. F. Stuessy (Eds.), Cladistics:
Perspectives on the reconstruction of evolutionary history (pp. 71-87).
New York: Columbia University Press.
Swofford, D. L. (1991). PAUP: Phylogenetic Analysis Using Parsimony.
Champaign: Illinois Natural History Survey
Phylogenetic Investigator 43
APPENDIX A -- MODEL PROBLEMS
Model Problems
The model problems were created to demonstrate fundamental concepts
in phylogenetic biology. In teaching, these can be useful both for modeling
problem-solving techniques and allowing students to practice recognizing these
patterns in the data. Each time a problem is selected, the taxa and characters
are randomly arranged, but the form of the solution will remain constant.
A solved example is provided for each of the model problems below with a
description of the number of taxa, characters, solutions (both topologies
(arrangements of taxa) and optimizations (character interpretations), and steps
(number of unweighted character transitions). For some problems, there are
also comments indicating particular features of interest.
These problems are also available separately in the "Model Problems" document.
Phylogenetic Investigator 44
Autapomorphy
2 taxa, 1 character, 1 solution with 1 step
Time
0
R85
10
20
1 O>l
30
40
F97
50
Morphological Change
1 Steps
1
R85 l
F97 O
Characters
2
3
4
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
T
a
x
a
Problem: Autapomorphy
Figure 27.
Autapomorphy: a phylogenetic tree representing an autapomorphy.
This problem demonstrates the essence of the phylogenetic problem: A taxon
at one point in time is ancestral (F99) with respect to a character of interest
(1) while a recent taxon (R84) has the character in the derived state. The
problem can be resolved by establishing a link of ancestral-descendant
relationship and placing a transition for the character on the link.
Phylogenetic Investigator 45
Synapomorphy 1
2 taxa, 1 character, 1 solution with 1 step
Time
0
Time
0
R81
10
R83
1 O>l
R81
R83
10
1 O>l
PA
1 O>l
20
20
F92
F92
30
30
40
40
50
50
Morphological Change
2 Steps
1
R81 l
R83 l
F92 O
Characters
2
3
4
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
Morphological Change
1 Steps
1
R81 l
R83 l
F92 O
Characters
2
3
4
5
6
7
T
a
x
a
T
a
x
a
Problem: Synapomorphy 1
Problem: Synapomorphy 1
Figure 28.
8 9 10 11 12 13 14 15 16 17 18 19 20
Synapomorphy 1: The data matrix contains a single character shared in
the derived state by two recent taxa. On the right this data is represented
as two autapomorphies (2 steps). More parsimonious is the phylogenetic
tree on the left representing a 2 taxon synapomorphy (1 step).
This problem demonstrates the fundamental assumption of modern
phylogenetic tree construction (What is sometimes called 'the auxiliary rule').
The two recent taxa share a derived character which is ancestral in the root. A
common ancestor can be postulated, linked to the recent taxa and to the oldest
taxon, and the transition for the character can be placed prior to the common
ancestor. For classroom modeling, it is often useful to initially construct this
problem as a convergence (both taxa linked directed to the ancestral taxon with
the transition occurring on each link) and then to reconstruct the problem
(using reassign links) to show synapomorphy. The principle of parsimony can be
introduced at this point.
Phylogenetic Investigator 46
Synapomorphy 2
3 taxa, 2 characters, 1 solution with 2 steps
Time
0
R89
R85
R84
10
PB
2 O>l
20
PA
30
1 O>l
40
F97
50
Morphological Change
2 Steps
1
R85 l
R89 l
R84 l
F97 O
Characters
2
l
O
l
O
3
4
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
T
a
x
a
Problem: Synapomorphy 2
Figure 29.
Synapomorphy 2: A phylogenetic tree representing a 3 taxon nested
synapomorphy.
This problem illustrates nested characters (What is sometimes called 'the
inclusion rule'). Characters 1 and 2 are nested because character 1's
distribution is included entirely within character 2's distribution. Being nested is
one way that characters can be 'consistent' or 'compatible.' Nested characters
represent a stronger hypothesis than exclusive characters.
Phylogenetic Investigator 47
Synapomorphy 3
4 taxa, 3 characters, 1 solution with 3 steps
Time
0
R80
R86
R84
R82
10
PA
20
PB
2 O>l
1 O>l
30
PC
3 O>l
40
50
F98
Morphological Change
3 Steps
1
R80 l
R82 O
R86 l
R84 O
T F98 O
Characters
2
O
l
O
l
O
3
l
l
l
l
O
4
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
a
x
a
Problem: Synapomorphy 3
Figure 30.
Synapomorphy 3: two 2-taxon synapomorphies joined by a whole ingroup
synapomorphy.
This problem illustrates mutually exclusive characters (What is sometimes called
'the exclusion rule'). Characters 2 and 3 are exclusive because their
distributions do not overlap. Exclusive characters, like nested characters, are
'compatible' or 'consistent' with one another.
Phylogenetic Investigator 48
Homoplasy 1 & 2
3 taxa, 4 characters, 1 topology and 2 optimizations, 4 steps
Time
0
Time
0
R85
R88
R80
R85
R88
1 l>O
10
R80
1 O>l
10
1 O>l
PB
PB
2 O>l
4 O>l
20
2 O>l
4 O>l
20
PA
PA
30
30
1 O>l
3 O>l
40
3 O>l
40
50
50
F99
F99
Morphological Change
5 Steps
1
R88 l
R80 O
R85 l
F99 O
Characters
2
l
l
O
O
3
l
l
l
O
4
l
l
O
O
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
Morphological Change
5 Steps
1
R88 l
R80 O
R85 l
F99 O
Characters
2
l
l
O
O
3
l
l
l
O
4
l
l
O
O
5
6
7
T
a
x
a
T
a
x
a
Problem: Homoplasy 1 & 2
Problem: Homoplasy 1 & 2
Figure 31.
8 9 10 11 12 13 14 15 16 17 18 19 20
Homoplasy 1 &2: A nested synapomorphy problem with 2 equally
parsimonious character optimizations..
This problem illustrates multiple character optimizations (convergence or
reversal). One character (3) conflicts with two other characters (1,2) resulting
in two different interpretations of the conflicting character. In one
interpretation, the conflicting character is gained twice. In the other, it is gained
once (before PA) and lost once (in R86)
Phylogenetic Investigator 49
Homoplasy 3
4 taxa, 4 characters, 1 solution 5 steps
Time
0
R80
R84
R87
R88
2 O>l
2 O>l
10
PC
PB
3 O>l
1 O>l
20
PA
4 O>l
30
F94
40
50
Morphological Change
5 Steps
1
R84 l
R87 O
R80 l
R88 O
T F94 O
Characters
2
l
l
O
O
O
3
O
l
O
l
O
4
l
l
l
l
O
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
a
x
a
Problem: Homoplasy 3
Figure 32.
Homoplasy 3: A convergence problem.
This problem illustrates homoplasy with a single resolution (convergence). Two
characters (1 and 3) are compatible and exclusive and nested within character
2. Character 4 conflicts with characters 1 and 3, but only one interpretation is
possible in this case. Constructing this solution as a reversal, would require two
reversals and, therefore, not be most parsimonious.
Phylogenetic Investigator 50
Homoplasy 4
3 taxa, 3 characters, 2 topologies each with 2 optimizations, 4 steps
Time
0
Time
0
R80
R83
R85
R85
R83
3 O>l
R80
2 O>l
10
10
3 O>l
2 O>l
PB
PB
2 O>l
3 O>l
20
20
PA
PA
30
30
1 O>l
1 O>l
40
40
F96
F96
50
50
Morphological Change
4 Steps
1
R80 l
R83 l
R85 l
F96 O
Characters
2
O
l
l
O
3
l
l
O
O
4
5
6
7
8 9 10 11 12 13 14 15 16 17 18 19 20
Morphological Change
4 Steps
1
R80 l
R83 l
R85 l
F96 O
Characters
2
O
l
l
O
3
l
l
O
O
4
5
6
T
a
x
a
T
a
x
a
Problem: Homoplasy 4
Problem: Homoplasy 4
Figure 33.
7
8 9 10 11 12 13 14 15 16 17 18 19 20
Homoplasy 4: A nested synapomorphy problem with 2 equally
parsimonious topologies each with 2 character optimizations (not
shown).
This problem illustrates multiple topologies. It is similar to Homoplasy 1 & 2,
except that there are now only two characters (1 and 2) that conflict with each
other. In the Homoplasy 1 & 2, having two identical characters unambiguously
defines the tree's structure. In this case, either character is equally believable
resulting in two arrangements of the taxa each with two character
interpretations (only the convergence optimization is shown for each topology).
Phylogenetic Investigator 51
APPENDIX B -- INSECT WING DATA SOURCE
Figure 34. The complete data source from which the insect wing example is selected.
The insect wings are also available as a PICT file, "Insect Wings.pict".
Phylogenetic Investigator 52
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement