Page 11 - iCC044-201501_March

An account of the outgroups used in analy-
sis is provided in the electronic supple-
mentary material. In some cases, there is
disagreement surrounding the identity of
the earliest representatives of these line-
ages. To accommodate uncertainty, two
sets of calculations were completed: one
using the oldest proposed minimum age
for a clade and the other applying the
youngest. Collectively, these paired analy-
ses provide upper and lower estimates
of CIs for divergence times given present
understanding of both the fossil record and
teleost interrelationships. These age esti-
mates are conditioned on a prior assump-
tion that divergence occurred after a user-
specified hard upper bound. This bound
applies to the divergences of all groups
considered, not only the focal clade.
We have therefore selected the Carbon-
iferous (Serpukhovian) Discoserra, a puta-
tive stem neopterygian [22], as defining an
upper bound of 322.8 Ma (see the elec-
tronic supplementary material).
(c) Collection of sequence data, phylogenetic a random local molecular-clock model in
analyses and relaxed molecular clocks
Standard phenol-chloroform extraction
protocol or Qiagen DNeasy Blood and
Tissue kits were used to isolate DNA from
tissue biopsies sampled from 158 spe-
cies of percomorph teleosts that included
89 species of Cichlidae (electronic sup-
plementary material, table S6). Previously
published PCR primers (see the elec-
tronic supplementary material) were used
to amplify and sequence exons from 10
nuclear genes (ENC1, Glyt, myh6, plagl2,
Ptr, rag1, SH3PX3, sreb2, tbr1 and zic1).
Amplified gene copies were cleaned and
used as templates for DNA cycle sequenc-
cichlidpower.org.au
Table 1. Posterior molecular age estimates for major lineages of Cichlidae.
Ages refer to crown groups.
Clade
Cichlidae
Etroplinae (India and Madagascar)
Ptychochrominae (Madagascar)
unnamed Afro-American clade
Cichlinae (neotropics)
Pseudocrenilabrinae (Africa)
unnamed east African clade
most recent common ancestor of L. Malawi and L. Victoria radiations
Crater Lake Barombi Mbo (Cameroon)
ing. Alignments of the DNA sequences
from the individual genes were constructed
from the inferred amino acid sequenc-
es. Thirty data partitions were designated
that corresponded to the three separate
codon positions for each of the 10 protein-
coding genes. A phylogeny of the aligned
DNA matrix was inferred using maximum-
likelihood and relaxed-clock analyses using
the computer program BEAST v. 1.6 (figure
1; electronic supplementary material, figure estimate for the fossil, outline the prior
S1) [23,24]. DNA sequences are depos-
ited on GenBank KF556709-KF557487.
Aligned gene sequences used in phyloge-
netic analyses, phylogenetic trees result-
ing from RAXML and BEAST analyses,
files formatted for BEAST analyses and
files used to estimate the age of cichlids
using palaeontological data are available
from the dryad digital repository (dx.doi.
org/10.5061/dryad.48f62). Fossil-based
age constraints were applied to 10 nodes
in the percomorph phylogeny (electronic
supplementary material, figure S2).
mean age (Ma) 95% highest posterior
density interval (Ma)
64.9 57.3-76.0
36.0 30.3-42.2
38.2 31.7-46.4
46.4 40.9-54.9
29.2 25.5-34.8
43.7 38.2-51.6
8.0 6.9-9.5
2.3 1.7-3.1
1.4 0.8-2.3
(d) Fossil calibration age priors
For each fossil calibration prior, we iden-
tify the calibrated node in the percomorph
phylogeny, list the taxa that represent the
first occurrence of the lineage in the fos-
sil record, describe the character states
that justify the phylogenetic placement of
the fossil taxon, provide information on
the stratigraphy of the rock formation(s)
bearing the fossil, give the absolute age
age setting in the BEAST relaxed-clock
analysis and provide any additional notes
on the calibration [25]. Each calibration is
numbered and the phylogenetic placement
of the calibration is highlighted in the elec-
tronic supplementary material, figure S3.
Full justification of our calibrations is given
in the electronic supplementary material.
Because we look to provide a critical test of
competing models of cichlid biogeography,
we have not assumed Gondwanan vicari-
ance a priori and did not use the timing of
the fragmentation history of this supercon-
tinent to inform calibrations in the relaxed-
9