suggests a more recent origin for the group (Pawlowski et al. 1999), which probably
evolved from filose protists belonging to the clade Cercozoa (Keeling 2001). The close
relationship between Foraminifera and cercozoans is also confirmed by analysis of SSU
sequences in the absence of fast evolving lineages (unpublished data; see Figure 6.1).
The Foraminifera are not the only example of misplacement in rRNA-based phylogenies
of protists, the best known being the case of Microsporidia, first considered to be very
primitive Eukaryotes (Vossbrinck et al. 1987), and now regarded as highly derived,
secondarily amitochondriate members of the Fungi (Hirt et al. 1999; Keeling et al. 2000).
However, the lack of a fossil record for this group makes it very difficult to determine
whether their rapid evolution is due to an episodic change of substitution rates or to other
factors. The same uncertainty exists for most other protists exhibiting high rates of
substitution, as groups with a satisfactory fossil record are unfortunately very scarce.
The rate acceleration in the foraminiferan stem-lineage also provides an example of the
kind of bias that can be introduced in the estimation of divergence times. If we calculate
the time of divergence between the Foraminifera and other Eukaryotes using the rates
estimated for different foraminiferan lineages, we will obtain the absurd value of more
than 9 Ga. In general, any episodic acceleration in a stem-lineage will lead to
overestimating the divergence time by the same factor for all groups belonging to this
lineage, each time the calibration is done using the sister group. Curiously, the variation
of rates in stem-lineages is usually not taken into account in the methods that have been
developed to estimate divergence times (see for instance Rambaut and Bromham 1998).
The Foraminifera and the Diptera are certainly not the only lineage of Eukaryotes
where such acceleration-deceleration patterns in the substitution rates of rRNA genes
might be observed, although in most cases it is probably not as striking—and not as easy
to reveal—as it is with Foraminifera. Several groups of organisms can be found at all
taxonomic levels, whose stem-lineage appears as a very long branch in rRNA-based trees,
and in which episodic rate changes are very likely to have happened. These include for
example the bilaterian metazoans (Giribet and Wheeler 1999), the florideophycean red
algae (Ragan et al. 1994), Plasmodium (Van de Peer and De Wachter 1997), and of course
even the Eukaryotes as a whole (Winker and Woese 1991; Morin 2000). Nevertheless,
the putative episodic nature of the observed differences in substitution rates of SSU rRNA
genes is generally not discussed in these studies.
Today, it is a common practice to discard the fast evolving species or lineages from
phylogenetic analyses and estimations of divergence times. Such a radical process merely
avoids the problem, rather than solving it. We think that a better understanding of the
patterns of molecular rate variations, including their episodic character, may be more
useful for accurate prediction of evolutionary origins and divergences.
AcknowledgementsThe authors thank Maria Holzmann, Colomban de Vargas, and José Fahrni for co-
operation in molecular data collecting. This study was supported by Swiss National
Science Foundation grant 31–59145.99.
EVOLUTION IN THE STEM-LINEAGE OF FORAMINIFERA 117