Biology Reference
In-Depth Information
divergence after gene transfer between species.
35
Such molecular homol-
ogy is probably the only type that has been well formalized in bioinfor-
matics up to now.
Evo-devo is dependent on definitions of homology both for anatom-
ical structures and for genes, and this field has also brought important
new information relative to homology. Some of this new information has
raised new questions, which are best exemplified by the now-classical case
of animal eyes — if insect eyes and vertebrate eyes are organized in fun-
damentally different ways, they are probably analogs; but if they are deter-
mined during development by orthologous genes, are they not homologs?
Such cases have led to the controversial proposal that the presence of key
orthologous genes in development suffices to define homology (discussed
in Wray and Abouheif
72
). An attempt to solve this issue is the new term
“homocracy”, which is defined as sharing the expression of the same pat-
terning genes.
73
Homocratic structures may or may not be homologous;
homologous structures are often homocratic, but this is not a logical
necessity.
When defining homology in development, we must also take into
account heterochrony, which is defined as changes in the relative timing
of development during evolution. For each organ, homology should be
defined at specific developmental stages, which differ between organs.
For example, the heart develops later in primates relative to rodents,
while the ear develops earlier, thus changing the timing of development
of these organs relative to each other.
74
This makes it impossible to
define homology between developmental stages as a whole in many
cases, and greatly complicates automatic comparison of embryos
between species. Moreover, developmental “stages” as defined in the lit-
erature are somewhat arbitrary divisions of a continuous process. The
limits of these divisions may not be consistent between species, even
without heterochrony.
Despite the importance of homology, little attention has been paid to
its careful implementation in bioinformatics. Notably, few ontologies
contain any notion of homology. Ontologies are formal representations
of a field of knowledge, including terms, definitions and relations
between the terms (e.g. hydrolase “is_an” enzyme). They have become
an important tool in bioinformatics, responding to the need to formalize
