ELECTRONIC SUPPLEMENTARY MATERIAL
Matrilineal evidence for demographic expansion, low diversity and lack of phylogeographic
structure in the Atlantic forest endemic Greenish Schiffornis Schiffornis virescens (Aves:
Tityridae)
Journal of Ornithology
Cabanne G Sa,b,c, Sari E Rd, Meyer Da, Santos F Rd, Miyaki CYa
aDepartamento
de Genética e Biologia Evolutiva, Instituto de Biociências, Universidade de São
Paulo, Rua do Matão 277, 05508–090, São Paulo, SP, Brazil.
bCONICET,
c División
Av. Rivadavia 1917, Ciudad de Buenos Aires (C1033AAJ), Argentina.
de Ornitologia, Museo Argentino de Ciencias Naturales “B. Rivadavia”, Ángel Gallardo
470, Ciudad de Buenos Aires (C1405DJR), Argentina.
dDepartamento
de Biologia Geral, Instituto de Ciências Biológicas, Universidade Federal de Minas
Gerais, Av. Antônio Carlos 6627, 31270-901, Belo Horizonte, MG, Brazil.
Supplementary material
Material and Methods
Coalescent simulations
In order to evaluate different demographic histories we simulated possible demographic scenarios in
the program BAYESSC, a modification of the software SERIAL SIMCOAL (Anderson et al. 2005;
Chan et al. 2006), and evaluated the goodness of fit of the observed data to the simulations. The
tested models (detailed in Figure 2) differed in the timing, number, and intensity of bottlenecks, and
also in the number of populations. The populations in model G were originated before the
Pleistocene, as suggested by the divergence of S. virescens with its sister species S. turdina (See Results)
In the simulations we organized the sampling localities (Table 1) by proximity into five
populations (pop); pop 1: localities 1-5; pop 2: locality 6-8, 11; pop 3: localities 12-14; pop 4: localities
9, 10, 15-17; pop 5: localities 18 and 19. To model panmixia we collapsed the five populations into
one at the first generation. Bottlenecks reduced Ne (effective size of genes) to 1-10% and 10-100% of
the present size. We obtained the species Ne from theta (Θ=2μNe) estimated in LAMARC 2.1.2b
(Kuhner 2006). Each population in the islands models, as well as the ancestral population of the
model G, presented an effective size equal to 1/5 of the species Ne. Estimations of Θ used the F84
model of sequence evolution, empirical base frequencies and transition/transversion ratios, with a
Markov chain Monte Carlo with default setting. Effective number of genes was introduced in
BAYESSC infiles as a normal distribution with mean equal to the maximum likelihood estimation of
Ne and a standard deviation estimated from Θ confidence interval. For the CR substitution rate we
used 1.67 x 10-8 substitution/site/MY (SE 1.67 x 10-9). We had previously obtained this value in
phylogenetic analyses in BEAST which used cytb and CR sequences and estimated the CR rate in
relation to a cytb rate (2.1 % divergence/MY, Weir and Schluter 2008). The transition bias was 0.79,
the mutation rate gamma distribution was 0.7576, and the number of mutation categories was six. We
assumed a generation time of one year.
For each model we ran 1,000 simulations. Then, for each simulated data we estimated
summary statistics for the complete data set to obtain null distributions against which we tested the
observed data. Summary statistics were: segregating sites, nucleotide diversity, Tajima’s D (Tajima
1989), Fu’s Fs (Fu 1997), and Φst (Excoffier et al. 1992). For evaluating the goodness of fit of the
observed data to simulated data 1) we used the two-tailed empirical likelihood pi of each summary


statistics i; pi  1  2 10.5  p (eq.1), being p the proportion of simulated values equal or higher
than the observed summary statistic. Then, 2) we combined the five pi values by obtaining
Cobs  2i1 ln pi (eq.2) and getting its significance. The significance was assessed by comparing
k
Cobs against a null distribution of C obtained according to Voight (2005) and Fabre (2009). Briefly, for
each simulated dataset, each value of summary statistic was compared with the other values
representing the empirical distribution of the statistic from simulation. Specifically, we treated the
value of each summary statistic as the observed value and calculated with eq. 1 its psim-value relative to
the remaining 999 simulated data. Then, the null distribution of C was obtained by combining with
eq. 2 values of psim across summary statistics and the significance of Cobs was obtained as in step 1
(eq.1). This procedure generated two-tailed global p-values associated to each model that we used to
evaluate plausibility of models.
REFERENCES
Anderson CNK, Ramakrishnan U, Chan YL, Hadly EA (2005) Serial SimCoal: A
population genetics model for data from multiple populations and points in time.
Bioinformatics 21 (8):1733-1734. doi:10.1093/bioinformatics/bti154
Chan YL, Anderson CNK, Hadly EA (2006) Bayesian estimation of the timing and
severity of a population bottleneck from ancient DNA. Plos Genetics 2 (4):451460. doi:e59
10.1371/journal.pgen.0020059
Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular variance inferred
from metric distances among DNA haplotypes - application to human
mitochondrial-DNA restriction data. Genetics 131 (2):479-491
Fabre V, Condemi S, Degioanni A (2009) Genetic Evidence of Geographical Groups
among Neanderthals. PLoS ONE 4 (4):e5151
Fu YX (1997) Statistical tests of neutrality of mutations against population growth,
hitchhiking and background selection. Genetics 147 (2):915-925
Kuhner MK (2006) LAMARC 2.0: maximum likelihood and Bayesian estimation of
population parameters. Bioinformatics 22 (6):768-770
Tajima F (1989) Statistical-method for testing the neutral mutation hypothesis by DNA
polymorphism. Genetics 123 (3):585-595
Voight BF, Adams AM, Frisse LA, Qian Y, Hudson RR, Di Rienzo A (2005)
Interrogating multiple aspects of variation in a full resequencing data set to infer
human population size changes. Proc Natl Acad Sci U S A 102 (51):18508-18513.
doi:10.1073/pnas.0507325102
Weir JT, Schluter D (2008) Calibrating the avian molecular clock. Mol Ecol 17
(10):2321-2328