AMAZON RIVER DOLPHIN: HIGH PHYLOPATRY DUE TO RESTRICTED DISPERSION AT LARGE AND SHORT DISTANCES - Biology, Evolution and Conservation of River Dolphins Within South America and Asia - page 104

30 s, and extension at 60 o C for 1 min. Resulting fragments were sequenced in a MegaBACE
1000 DNA Analysis System with protocols indicated by the fabricant (GE Healthcare®).

Molecular Data and Statistic Analysis

Sequences of the mtDNA control region (HVSI) were ―base called‖ with the software
Phred 0.020425.c (Ewing et al., 1998). High quality consensus sequences (519 bp for HVSI
and 1140 bp for Cyt-b) for each individual were produced from alignments of forward and
reverse strand sequences with the software Phrap 0.990319 (Green, 1994) and
visualized/edited in Consed 12.0 (Gordon et al ., 1998). The consensus sequences from all
individuals were aligned using the Clustal-X 1.83 (Thompson et al., 1997) to allow the
identification of nucleotide variation and the different haplotypes. All control regions mtDNA
sequences were deposited in the Genbank with the accession numbers DQ217403 to
DQ217408 and EU55456/EU554562 for the Cyt-b sequences. In addition to our data, we
included in the analyses all HVSI and Cyt-b sequences of I. geoffrensis specimens published
elsewhere (Cassens et al., 2000; Banguera-Hinestroza et al., 2002; Arnasson et al., 2004).
Phylogenetic analyses with HVSI mtDNA were carried out using the program Mega 4
(Kumar et al., 2004) using the neighbor-joining method with Tamura-Nei (TN - Tamura &
Nei, 1993) substitution model, and gamma distributed rates among sites ( = 0.88) as
suggested by the Modeltest approach (Posada & Crandall, 1998). For Cyt-b data, we used the
neighbor-joining algorithm and Kimura 2-parameters genetic distance with ( = 0.5. Branch
support tests were calculated by bootstrap using 1000 replicates with Mega. The sequence of
Pontoporia blainvillei (Genbank accession number AY644451) was used as the outgroup for
the phylogenetic analyses.
Haplotype networks were constructed by the median-joining algorithm (Bandelt et al.,
1999) using the program Network v 4.5 (www.fluxusengineering.com) to infer the
relationship among the haplotypes for the populations in a geographic context.
Population genetics parameters such as haplotype diversity ( h ) and nucleotide diversity
() were estimated using either the average number of pairwise differences (K) or alternative
models of evolution (see above) in the Arlequin 3.1 software (Schneider et al., 2000). We
used Mega to calculate average sequence pairwise divergences within and between
populations using TN + Γ ( = 0.88) model.
The analysis of molecular variance (AMOVA), a test to calculate the distribution of
genetic variation in a particular hierarchical grouping of populations, was performed with
1,000 permutations, using either F ST or  ST calculated with TN + Γ ( = 0.88) in Arlequin.
Tajima's D-test (Tajima, 1989) and Fu's Fs test (Fu, 1997) were calculated for each
population to evaluate the possibility of recent population expansion using 1,000 bootstrap
replicates. Mismatch distributions (Schneider et al., 2000) were also used to evaluate the past
occurrence of population bottlenecks and expansion. An exact test of population
differentiation was performed with 10,000 Markov Chain steps to test significance as
implemented in Arlequin. Population pairwise distances were calculated as F ST and  ST (TN,
=0.88). The number of female migrants between populations was estimated as

1 following Slatkin's (1991) formula. Mantel tests were performed to test the  

mf N 



ST

2

ST