Colour pattern as a single trait driving speciation in Hypoplectrus coral reef fishes?
Electronic Supplementary Material
Oscar Puebla, Eldredge Bermingham, Frédéric Guichard and Elizabeth Whiteman
Supplementary Methods
Genetics analyses
Identification, catalogue number, collection sites and collection dates of all samples are detailed in Supplementary Table 3. Total DNA was extracted using column DNeasy® Tissue Kits (Quiagen). An enriched Hypoplectrus microsatellite library was built following Kandpal et al.(1994) and 5 microsatellite markers were developed (GenBank accession no. DQ517318, DQ517319, DQ517320, DQ517321, DQ517322,Supplementary Table 4). These markers as well as 5previously existing ones(McCartney et al. 2003) were tested and optimized on 84 individuals from 7 colour morphs (Supplementary Table 4). PCR amplifications were performed on PTC-200 Peltier Thermal Cyclers (MJ Research) in a 12-ml final volume containing 10-100 ng DNA, 0.2 mM each dNTP, 0.01 mM primer F, 0.4 mM primer R, 0.15 mM primer M13, 1x PCR Buffer II with no MgCl2 (Applied Biosystems, unknown composition), 2 mM MgCl2, and 0.25 U of AmpliTaq® DNA polymerase (Applied Biosystems). All primers F contained a 5’ M13 extension and the additional M13 primer (5’cacgacgttgtaaaacgac3’) was fluorescently labelled at 5’ with 6-FAM, VIC, NED, or PET (Applied Biosystems; Supplementary Table 4). PCR cycling parameters were as follow: 40s of initial denaturing at 92°C; 30 cycles of 40s at 92°C, 40s at annealing temperature (Supplementary Table 4), 40s at 72°C; and a final extension of 3 min at 72°C. Amplified fragments were separated by capillary electrophoresis on an ABI PRISM® 3130xl automated genetic analyser (Applied Biosystems) using 500 LIZ® as size standard and analysed with genescan and genotyper software packages (Applied Biosystems).
Statistical analyses of genetic data
Linkage disequilibrium between loci was tested with genotype randomizations overall samples between all pairs of loci (4,500 permutations) using the log-likelihood ratio G-statistic with FSTAT version 2.9.3(Goudet 1995). Hardy-Weinberg equilibrium was tested with allele randomizations within samples (1,000 permutations per test) and overall samples (10,000 permutations) with FSTAT. The statistics used for these tests were Fis and Fit for within-samples an overall test respectively, both estimated following Weir & Cockerham(1984).
An allele size permutation test(Hardy et al. 2003) was performed using SPAGeDi version1.2(Hardy & Vekemans 2002) to test whether stepwise-like mutations contributed significantly to genetic differentiation (10,000 allele size permutations per locus). Stepwise-like mutations did not contribute significantly to genetic differentiation in 9 loci out of the 10 analysed (p-values between 0.156 and 0.973). Fst rather than Rst(Slatkin 1995) were therefore considered for the following analyses. Fst between sympatric colour morphs was estimated following Weir & Cockerham (1984) with FSTAT. Genetic differentiation was tested with 10,000 allele permutations using the G log-likelihood statistic with FSTAT as this test proved to be more efficient than Fst-estimator based tests when Fst values are low(Goudet et al. 1996).
Supplementary Table 3. Identification, catalogue number, collection sites, and collection dates of all samples considered in this study.
Continued
Continued
Continued
Continued
Supplementary Table 4. Description of 10 microsatellite markers tested on 84 individuals from 7 Hypoplectrus colour morphs. n=12 per colour morph, gene diversities estimated following Nei(1987).
1Ball et al (unpublished), 2Chapman et al (1999), 3McCartney et al (2003), 4this study, Ta: annealing temperature
Supplementary Table 5. Fis estimate(Weir & Cockerham1984), heterozygote deficit test p-value and heterozygote excess test p-value per locus and sample. All loci are at Hardy-Weinberg equilibrium at all samples.
n: number of individuals analysed per sample.
Supplementary Discussion
Colour pattern
Seven colour morphs (H. nigricans, H. puella, H. unicolor, H. indigo, H. aberrans, H. guttavarius, and tan hamlet)were observed in Belize, 6 (H. puella, H. nigricans, H. unicolor, H. aberrans, H. gummigutta, and tan hamlet) in Bocas del Toro, 6 (H. puella, H. nigricans, H. unicolor, H. aberrans, H. indigo, and tan hamlet) in Kuna Yala, and 4 (H. puella, H. nigricans, H. chlorurus, and H. guttavarius) in Barbados. H. puella and H. nigricans were the only colour morphs present at all study sites.
Apart from the rare occurrence of “intermediate” individuals, colour pattern was clearly discrete at each site and most samples could be attributed to one colour morph confidently. That sympatric colour morphs constitute discrete phenotypic entities is generally not disputed in the Hypoplectrus literature, even by authors who regard colour morphs as polymorphs of a single species(e.g. Graves & Rosenblatt 1980). Note that the Kuna indigenous people from Panama recognize and differentiate the 5 colour morphs present in Kuna Yala as well.
Genetics
The 10 microsatellite markers amplified successfully and were polymorphic in the 7 colour morphs tested (Supplementary Table 4). The 371 samples considered in this study were genotyped at the 10 microsatellite loci with less than 1.4% missing data (locus poorly or not amplifying for a given sample) overall. A total of 231 alleles were identified, 138 of which where shared by the 3 colour morphs examined. Forty-two alleles were observed in one colour morph exclusively yet all were rare, with a frequency smaller than 1% overall samples for the locus considered.
There was no evidence of linkage disequilibrium between any pair of loci at the 0.05 level after sequencial Bonferroni correction for multiples tests (Rice 1989). All loci were at Hardy-Weinberg equilibrium within all samples after sequencial Bonferroni correction for multiple tests (Supplementary Table 5). Even without corrections for multiple tests, only 5 departures from Hardy-Weinberg would be significant (3 heterozygote deficits and 2 excesses) out of the 160 tests performed (heterozygote deficit and excess on 10 loci x 8 samples, Supplementary Table 5). In contrast 6 loci out of the 10 analysed presented a deficit in heterozygotes overall samples (p-values between 0.0008 and 0.0372) and there was a global deficit in heterozygotes overall samples and loci (Fit estimate=0.033, p-value<0.0001), suggesting reduced gene flow between samples.
Supplementary References
Chapman, R.W., Sedberry, G.R., Koenig, C.C. & Eleby B.M. 1999 Stock identification of gag, Mycteroperca microlepis, along the southeast coast of the United States. Marine Biotechnol.1, 137-146.
Goudet, J. 1995 FSTAT (version 1.2): a computer program to calculate F-statistics. J. Hered.86, 485-486.
Hardy, O.J. & Vekemans X. 2002 SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Mol. Ecol. Notes2, 618-620.
Hardy, O. J., Charbonne N., Fréville, H. & Heuertz, M. 2003 Microsatellite allele sizes: a simple test to assess their significance on genetic differentiation. Genetics163, 1467-1482.
Kandpal, R. P., Kandpal, G. & Weissman, S. M. 1994 Construction of libraries enriched for sequence repeats and jumping clones, and hybridization selection for region-specific markers. Proc. Natl. Acad. Sci. USA 91, 88-92.
Nei, M. 1987 Molecular Evolutionary Genetics. Columbia University Press: New York.
Rice, W.R. 1989 Analyzing tables of statistical tests. Evolution43, 223-225.
Slatkin, M. 1995 A mesure of population subdivision based on microsatellite allele frequencies. Genetics139, 457-462.
Supplementary Video Legends
Supplementary Video Sequence 1. Illustration of the behavioural pattern considered as “tracking” in the present study. Putative mimic H. unicolor presents clear changes in speed and/or direction in order to stay within approximately 30 cm of two C. capistratus putative models. This sequence would be considered as 24 seconds of “tracking” in our study.
Supplementary Video Sequence 2. Illustration of the behavioural pattern considered as “predatory strike” in the present study. H. unicolor performs a clear, sharp, and long acceleration. Note that the strike is performed close to a C. capistratus and that the prey, a small fish, is visible in this sequence. This sequence would be considered as one “predatory strike” in the presence of C. capistratus in our study.
1