Computational protocol: River barriers and cryptic biodiversity in an evolutionary museum

Similar protocols

Protocol publication

[…] Our field sampling was conducted in 2010 and 2011. In 2010, we sampled both north and south of the Congo River, near Kisangani (). In 2011, we again sampled north and south of the Congo River, but samples also were collected on opposite sides of Congo River tributaries to increase potential across-river comparisons. On the north bank, we collected west of the Lindi River, whereas on the south bank, we collected west of the Lomami River. Birds were collected via mist nets and euthanized following approved protocols (Texas A&M University Animal Use Protocol No. 2009-028). Collection sites and bird species included in this study are presented in ; the variation in numbers across species is indicative of their population densities. For both Bleda syndactyla and Stiphrornis xanthogaster, we also included preserved tissues samples collected by the Field Museum of Natural History (FMNH) from areas north and well to the east of the Congo River (triangles in ).Birds were prepared in the field as standard museum skins, with associated tissue and blood samples. Prior to preparation, birds were exposed to ethyl acetate and brushed for louse ectoparasites. Birds were identified to species and lice to genus using morphology and both have been catalogued into the Texas Cooperative Wildlife Collection at Texas A&M University. Of the specimens brushed for ectoparasites, there were only three instances where lice were collected from individuals of the same bird species on opposite sides of the Congo River: Myrsidea sp. (Phthiraptera: Amblycera) parasitizing Alethe castanea, Myrsidea sp. parasitizing Terpisphone rufiventer and Sturnidoecus sp. (Phthiraptera: Ischnocera) parasitizing Terpsiphone rufiventer.Our strategy was to assess whether haplotypes on opposite sides of a river(s) were reciprocally monophyletic with respect to one another and whether the pattern (if any) across species was temporally similar, indicating a shared evolutionary history with respect to divergence time.For birds and lice, whole genomic DNA was extracted from tissue using the DNeasy tissue extraction kit (Qiagen Inc., Valencia, CA); louse specific protocols were used for lice (Cruickshank et al. ; Johnson and Clayton ). After DNA extraction, lice were mounted on slides and retained as vouchers. We used the polymerase chain reaction (PCR) to amplify portions of the cytochrome-b (cyt-b) gene for all avian samples; we also amplified the nuclear βact3 gene for Bleda syndactyla and the nuclear β-fibrinogen intron-5 gene (TGFβ2) for Stiphrornis xanthogaster. We amplified the cytochrome oxidase I (COI) gene for lice. PCR amplifications used published primers and protocols (Sorenson et al. ; Outlaw et al. ; Carling and Brumfield ; Light and Reed ). Automated sequencing was performed using BigDye (Applied Biosystems, Carlsbad, CA) and products were run on an ABI 377 sequencer at the University of Florida ICBR facility. We used SEQUENCHER, version 4.5 (Gene Codes, Ann Arbor, MI) to align sequences. To ensure the accuracy of amplification, we sequenced both heavy and light strands, and verified that sequence data were protein-coding. Sequences for louse taxa and those bird species for which genetic variation was evident (Alethe, Bleda syndactyla, Illadopsis, Stiphrornis) have been deposited on GenBank under accession numbers KC349953-KC349959 (lice) and KC355099-KC355178 (birds).Uncorrected p-distances were examined for louse taxa and phylogenetic and other analyses were focused on avian taxa with genetic variation north and south of the Congo River (see Results: Alethe castanea, Bleda syndactlya, Illadopsis rufipennis, and Stiphrornis xanthogaster). We used MrModelTest (Nylander ) to determine the appropriate model of nucleotide substitution for each avian lineage (GTR+I, GTR+I+G, HKY+I, and HKY+G for Alethe, Bleda syndactyla, Illadopsis, and Stiphrornis, respectively). Sequence data for each understory species were analyzed using MRBAYES (Huelsenbeck and Ronquist ), where we initiated two runs of four Markov–chain Monte Carlo (MCMC) chains of 2 million generations each from a random starting tree, sampling every 100 generations. Each run resulted in 20,000 trees and converged on the same topology. The first 50,000 generations (5000 trees) from each analysis were removed as our “burn-in”, and the remaining 30,000 trees were used to create a majority rule consensus tree. Trees were rooted using mid-point rooting.We used the program BEAST v1.6.1 (Drummond and Rambaut ) to estimate TMRCA within each species. We employed a lineage substitution rate of 0.0105 per site/million years using a relaxed, uncorrelated lognormal clock. This substitution rate translates to 2.1% per million years, and is generally applicable to the cyt-b gene in songbirds (Weir and Schluter ; Lerner et al. ). We also employed a normal distribution for this prior and assigned a standard deviation of 0.0013, which encompasses a slower (1.6%) and faster (2.53%) estimate calculated for songbird cyt-b substitution rates in other studies (Fleischer et al. ; Nabholz et al. ). A Yule process speciation prior was implemented in each analysis. Two separate MCMC analyses were run for 10,000,000 generations with parameters sampled every 1000 steps, and a 10% burn-in. Independent runs were combined using LogCombiner v.1.6.1 (Drummond and Rambaut ; Drummond et al. ). Tracer v.1.5 (Rambaut and Drummond ) was used to measure the effective sample size of each and calculate the mean and upper and lower bounds of the 95% highest posterior density interval (95% HPD) for divergence times. Tree topologies were assessed using TreeAnnotator v.1.7 (Drummond and Rambaut ; Drummond et al. ) and FigTree v.1.3.1 (Rambaut ).*BEAST (Drummond and Rambaut ; Heled and Drummond ) was used to infer a “multi-species” tree (consisting of two taxa: “north” and “south” of the Congo River) from cyt-b gene trees obtained from Alethe, Bleda syndactlya, Illadopsis, and Stiphrornis. *BEAST was used to determine if the Congo River caused divergence within each of these bird species at approximately the same time. Specimens were identified as being collected from north or south of the Congo River and used to reconcile the evolutionary history of bird taxa in the Congo Basin. All analyses were run in BEAST v1.7 (Drummond and Rambaut ; Drummond et al. ) using the models of molecular evolution identified in the phylogenetic analyses (see above). Several analyses were run to reconcile the evolutionary history of these four species. Preliminary BEAST analyses resulted in the parameter ucld.stdev (the standard deviation of the uncorrelated lognormal relaxed clock) being close to 0 (suggesting that the avian cyt-b data are clocklike). We therefore performed analyses enforcing a strict clock as well as a relaxed, uncorrelated lognormal clock for the substitution rate for comparison. We also varied the speciation prior, using a Yule process as well as a speciation birth-death process speciation prior. Relative evolutionary rates were estimated among the three gene trees (Bleda syndactyla, Illadopsis, and Stiphrornis) by setting the prior of the rate for Alethe at 1.0. For each analysis, two separate MCMC runs were run for at least 60 million generations with parameters sampled every 1000 steps, and incorporating a 10% burn-in. Combining of MCMC runs, and assessment of parameters and trees were performed as above.MsBayes (Hickerson et al. , ) was used to test for simultaneous divergence times in Alethe, Bleda syndactyla, Illadopsis, and Stiphrornis. Individuals from each species were identified as being collected from north or south of the Congo River and data from all four species were considered in combination and number of divergences was evaluated. One million simulations were drawn from the hyper-prior and the hyper-posterior was constructed from 1000 samples (tolerance = 0.005) using the hierarchical approximate Bayesian computation acceptance/rejection algorithm. […]

Pipeline specifications