genes and nuclear rRNA in resolving deep divergences and providing rate estimates, they
add a level of complication due to gene duplications and gene loss (Nichols 2001). Gene
duplication can introduce marked differences between gene phylogenies and species
phylogenies, particularly in the vertebrates where many genes actually occur as large gene
families (Page 2000). Attempts to infer phylogeny from nuclear genes must take gene
duplications into account and ensure as far as possible that the sequences under
comparison are orthologous rather than paralogous. For example, one criterion for
ensuring that sequences are orthologous is to check that the phylogeny of the sequences
being used matches the known phylogeny of the species (Nei et al. 2001). However, this
may become too lax a criterion when the number of sequences is low. Alternatively,
‘blasting’ a sequence in Genbank (Altschul et al. 1997) to find genes with high sequence
similarity can provide information about the existence of orthologues in the species under
scrutiny.
Molecular evidence: criticisms and critical testsAll molecular estimates of protostome-deuterostome divergence times fall in the
Precambrian with the majority of studies ranging from 800–1000 Ma (Figure 3.3). The
‘individual protein’ (IP) approach has been taken in a number of metazoan divergence
studies (e.g. Kumar 1996; Ayala et al. 1998; Gu 1998). With this approach, individual
divergence time estimates are made for each protein, and the mean of these estimates is
often presented as the final estimate. Nei et al. (2001) suggested that this approach has an
Figure 3.3 The confidence intervals on times of divergence between protostome and deuterostome
phyla as measured by recent studies. Estimates are sorted in chronological order. Dots indicate
point estimates. Line thickness indicates number of genes analysed. Black lines, grey lines and the
ribosome image indicate nuclear, mitochondrial and 18S genes used in the analyses, respectively.
PHYLOGENETIC FUSES AND EVOLUTIONARY ‘EXPLOSIONS’ 49