5 Metazoan Complexity
145
This notion of a steady gain of “complexity” in evolution has remained an influential template for the interpretation and evaluation of biological data on many
different levels, ranging from genome characteristics to the number and characteristics of tissue types in an organism. The expectation that complexity steadily
increases over time (and the often implicit assumption that primates represent the
epitome of animal complexity) lets exceptions of this principle appear odd and
counterintuitive. For instance, prior to the discovery of non-coding DNA, the term
“c-value paradox” was used to describe the apparent lack of correlation between
genome sizes and perceived complexity of animals (reviewed in Gregory 2005). The
“primitive” lungfish Protopterus aethiopicus (Pedersen 1971) has a genome that is
about 400 times larger than that of the “highly evolved” teleost genome of the puffer
fish, Tetraodon nigroviridis (Jaillon et al. 2004), and about 40 times larger than the
human genome.
After the c-value paradox was resolved by the discovery that these changes result
mainly from the different amounts of non-coding DNA, the expectation was, in
keeping with the notion of increasing complexity, that the amount of coding DNA
(or the number of genes) would be higher in “complex” animals than in “simpler”
ones. Cases like the duplication of the Hox cluster during chordate evolution, and
the notion that this duplication was part of two rounds of whole-genome duplications (the “2R-hypothesis”) lent support to the general notion that gene duplications
were a major source of developmental innovation (reviewed in Taylor and Raes
2004). As we now know, however, the total excess of coding genes in the currently sequenced vertebrate genomes over those of the classical invertebrate models,
Caenorhabditis and Drosophila, is surprisingly slim (Claverie 2001), and different vertebrates have roughly similar gene numbers. Moreover, important aspects of
developmental patterning, such as the impact of the transcriptional regulator pax6
on photosensitive structures (Halder et al. 1995), or the role of Hox genes in anteroposterior patterning (McGinnis et al. 1984), date back to early evolutionary times,
providing evidence for an ancient core “developmental genetic toolkit” acting in
animal development (see Cañestro et al. 2007). Considering that the connections
between different components of this toolkit are of crucial importance for animal
development, it is clear that a thorough analysis of animal complexity requires more
than a simple quantification of genes or genetic functions.
On a more morphological level, animal complexity has been assessed by counting the number of morphologically distinct adult cell types (Bell 1997, Sempere
et al. 2006, Valentine et al. 1994), sometimes complemented by the presence or
absence of other morphological and embryonic characters (Aburomia et al. 2003,
Heimberg et al. 2008). This method, however, has several pitfalls: as for some
animals, more morphological data is available than for others, the complexity of
well-analysed animals tends to be overestimated (Bonner 1988). In addition, morphological features fall short of describing other dimensions of complexity, such as
the developmental and behavioural complexity of an animal, including life history
strategies (larval or direct development) or the ontogeny of cell types (Bonner 1988,
Valentine 2000, Valentine et al. 1994). Finally, morphological features are also difficult to quantify and weigh against each other, generating a need for more simple
and quantitative measures of complexity.
145
This notion of a steady gain of “complexity” in evolution has remained an influential template for the interpretation and evaluation of biological data on many
different levels, ranging from genome characteristics to the number and characteristics of tissue types in an organism. The expectation that complexity steadily
increases over time (and the often implicit assumption that primates represent the
epitome of animal complexity) lets exceptions of this principle appear odd and
counterintuitive. For instance, prior to the discovery of non-coding DNA, the term
“c-value paradox” was used to describe the apparent lack of correlation between
genome sizes and perceived complexity of animals (reviewed in Gregory 2005). The
“primitive” lungfish Protopterus aethiopicus (Pedersen 1971) has a genome that is
about 400 times larger than that of the “highly evolved” teleost genome of the puffer
fish, Tetraodon nigroviridis (Jaillon et al. 2004), and about 40 times larger than the
human genome.
After the c-value paradox was resolved by the discovery that these changes result
mainly from the different amounts of non-coding DNA, the expectation was, in
keeping with the notion of increasing complexity, that the amount of coding DNA
(or the number of genes) would be higher in “complex” animals than in “simpler”
ones. Cases like the duplication of the Hox cluster during chordate evolution, and
the notion that this duplication was part of two rounds of whole-genome duplications (the “2R-hypothesis”) lent support to the general notion that gene duplications
were a major source of developmental innovation (reviewed in Taylor and Raes
2004). As we now know, however, the total excess of coding genes in the currently sequenced vertebrate genomes over those of the classical invertebrate models,
Caenorhabditis and Drosophila, is surprisingly slim (Claverie 2001), and different vertebrates have roughly similar gene numbers. Moreover, important aspects of
developmental patterning, such as the impact of the transcriptional regulator pax6
on photosensitive structures (Halder et al. 1995), or the role of Hox genes in anteroposterior patterning (McGinnis et al. 1984), date back to early evolutionary times,
providing evidence for an ancient core “developmental genetic toolkit” acting in
animal development (see Cañestro et al. 2007). Considering that the connections
between different components of this toolkit are of crucial importance for animal
development, it is clear that a thorough analysis of animal complexity requires more
than a simple quantification of genes or genetic functions.
On a more morphological level, animal complexity has been assessed by counting the number of morphologically distinct adult cell types (Bell 1997, Sempere
et al. 2006, Valentine et al. 1994), sometimes complemented by the presence or
absence of other morphological and embryonic characters (Aburomia et al. 2003,
Heimberg et al. 2008). This method, however, has several pitfalls: as for some
animals, more morphological data is available than for others, the complexity of
well-analysed animals tends to be overestimated (Bonner 1988). In addition, morphological features fall short of describing other dimensions of complexity, such as
the developmental and behavioural complexity of an animal, including life history
strategies (larval or direct development) or the ontogeny of cell types (Bonner 1988,
Valentine 2000, Valentine et al. 1994). Finally, morphological features are also difficult to quantify and weigh against each other, generating a need for more simple
and quantitative measures of complexity.