Biology Reference
In-Depth Information
Purnell
62
), we need to understand its mechanisms as well as possible.
Gene retention may also be biased relative to function, and enrichments
in communication and developmental genes have been reported in
insects, yeasts,
63
and
Arabidopsis
.
64
In fish, we found an excess of genes
annotated with terms related to development and signaling functions,
57
which supports the putative link between genome duplication and devel-
opmental innovations.
Second, once the importance of biased gene retention after duplica-
tion is established, this provides a convenient measure of selective pres-
sures on the genome. We have used this to test developmental constraints
on genome evolution. Two models have been proposed for develop-
mental constraints on morphological evolution. The first, dating back to
pre-Darwinian observations by von Baer,
1,65
is that there is a progressive
divergence of morphological similarities between vertebrate embryos.
More general characters would form in early development, which would
be highly constrained; while species-specific characters would form in late
development, which would be more open to innovation. An alternative
model was proposed more recently
66,67
: the “hourglass model”, which is
based on the observation of large morphological diversity in very early
development (e.g. blastula). This model assumes a constrained stage in
middle development, around the vertebrate pharyngula stage. To evalu-
ate the impact of the constraints postulated by both models on the
genome, we investigated the pattern of expression during development
according to gene retention after whole-genome duplication. We expect
genes that are kept in double to be highly expressed at developmental
stages that are open to evolutionary innovation, but not at stages that are
highly constrained. These should be characterized by conservatism: no
duplication and no loss of highly expressed genes (Fig. 3). We also expect
a high cost of gene loss (e.g. lethal phenotype for gene KO) in more con-
strained stages.
By combining a zebrafish time series microarray experiment
(E-TABM-33 accession from ArrayExpress) with phylogenetic definition
of retention or loss after duplication in fish (Fig. 2), we find that genes
kept in duplicate are lowly expressed in early development, and then
increase regularly their expression to reach a maximum in late develop-
ment.
68
In principle, this could be the result of biased evolution after
