4 Phylogeny of Animals
121
in the development of unambiguously convergent organs such as limbs of arthropods and vertebrates (Shubin et al. 1997). Conversely, dissimilar body plans such as
those of cnidarians and vertebrates share the extensive genetic networks that underlie anteroposterior and dorsoventral patterning (Martindale 2005). As a whole, this
rephrases in molecular terms the problem of homology assignment that caused considerable problems in many morphological studies (Gould 2002, Wagner 2007). The
accurate interpretation of developmental and morphological characters thus requires
the use of an independent reference, which would help to determine the steps of animal evolution. The search for this reference has been pursued by the development
of molecular inference methods and the study of signatures within some specific
genes. This has led to a reevaluation of animal classification with the rise of the
“new view” of animal phylogeny (Halanych 2004), which is based extensively on
the use of 18S ribosomal RNA as a universal molecular marker. However, despite
its indisputable success, this approach failed to solve some long-standing issues of
animal phylogeny such as, for example, the resolution of relationships at the base
of the metazoan tree or the branching of some incertae sedis (e.g. chaetognaths or
acoel flatworms, Fig. 4.1). These limitations could be explained by two possible
problems: (a) stochastic error related to the insufficient amount of information in
the sampled marker genes and (b) systematic errors related to the impossibility for
the molecular evolution model to fully account for the data.
The availability of a growing amount of genomic data for a broad range of organisms is now opening up a new field of investigation and providing new clues about
animal relationships. These genomic data enable the assembly of datasets composed of a large number of protein coding genes. This can eliminate the problem
of stochastic error, but can also help to identify more qualitative molecular characters, sometimes called “rare genomic changes” (Rokas and Holland 2000a). In this
chapter, we will show how these genomic approaches have reshaped our view of
animal relationships by attempting to describe both the successes and the pitfalls
of these approaches. The development of improved inference methods but also the
sequencing effort undertaken have led to some deep rearrangements of metazoan
trees that will certainly change our view of animal evolution.
4.2 The Roots of Animal Phylogeny
Before the molecular biology era, the science of systematics had attempted to
deal with the morphological complexity by proposing various methods to handle
character interpretation.
4.2.1 Historical Schemes Are Based on the Coelom Evolution
Hypotheses
The “theory of recapitulation” formulated by Haeckel in 1866 postulated that
the successive steps of evolution had been accomplished through the addition of
supplementary embryological stages (Gould 1977). Thus, the deepest levels of
121
in the development of unambiguously convergent organs such as limbs of arthropods and vertebrates (Shubin et al. 1997). Conversely, dissimilar body plans such as
those of cnidarians and vertebrates share the extensive genetic networks that underlie anteroposterior and dorsoventral patterning (Martindale 2005). As a whole, this
rephrases in molecular terms the problem of homology assignment that caused considerable problems in many morphological studies (Gould 2002, Wagner 2007). The
accurate interpretation of developmental and morphological characters thus requires
the use of an independent reference, which would help to determine the steps of animal evolution. The search for this reference has been pursued by the development
of molecular inference methods and the study of signatures within some specific
genes. This has led to a reevaluation of animal classification with the rise of the
“new view” of animal phylogeny (Halanych 2004), which is based extensively on
the use of 18S ribosomal RNA as a universal molecular marker. However, despite
its indisputable success, this approach failed to solve some long-standing issues of
animal phylogeny such as, for example, the resolution of relationships at the base
of the metazoan tree or the branching of some incertae sedis (e.g. chaetognaths or
acoel flatworms, Fig. 4.1). These limitations could be explained by two possible
problems: (a) stochastic error related to the insufficient amount of information in
the sampled marker genes and (b) systematic errors related to the impossibility for
the molecular evolution model to fully account for the data.
The availability of a growing amount of genomic data for a broad range of organisms is now opening up a new field of investigation and providing new clues about
animal relationships. These genomic data enable the assembly of datasets composed of a large number of protein coding genes. This can eliminate the problem
of stochastic error, but can also help to identify more qualitative molecular characters, sometimes called “rare genomic changes” (Rokas and Holland 2000a). In this
chapter, we will show how these genomic approaches have reshaped our view of
animal relationships by attempting to describe both the successes and the pitfalls
of these approaches. The development of improved inference methods but also the
sequencing effort undertaken have led to some deep rearrangements of metazoan
trees that will certainly change our view of animal evolution.
4.2 The Roots of Animal Phylogeny
Before the molecular biology era, the science of systematics had attempted to
deal with the morphological complexity by proposing various methods to handle
character interpretation.
4.2.1 Historical Schemes Are Based on the Coelom Evolution
Hypotheses
The “theory of recapitulation” formulated by Haeckel in 1866 postulated that
the successive steps of evolution had been accomplished through the addition of
supplementary embryological stages (Gould 1977). Thus, the deepest levels of