Citation Context

...ibility. Even in the absence of compositional heterogeneity, rate heterogeneity among lineages can confound phylogenetic estimates (e.g., Felsenstein, 1978; Hasegawa et al., 1991; Steel et al., 1993; =-=Ho and Jermiin, 2004-=-), so there is a need for more general nucleotide substitution models that relax the assumptions of stationarity, reversibility, and homogeneity. Based on the maximum-likelihood criterion, such models...

Citation Context

...ese published studies. In this paper I highlight the potentially serious sources of error in this method, and include a more detailed discussion of some of the pitfalls pointed out by Ronquist. Because the biogeographic history of any taxon is unknown it is impossible to test whether diva or any other method has correctly reconstructed its history. Simulation studies in phylogenetics have proved extremely useful for testing the accuracy of tree-reconstruction methods in recovering the known, albeit hypothetical, phylogeny of the simulated sequences (Poe & Swofford, 1999; Holland et al., 2003; Ho & Jermiin, 2004). Here, I use a parallel approach to highlight circumstances under which inferences based on diva can be wrong. Hypothetical taxa are allowed to evolve in a specific geological scenario, and diva is used to estimate the ancestral areas. SCENARIO Assume four landmasses, A, B, C and D, that are connected to each other at some point in time. These landmasses separate following the advent of barriers between them (e.g. ocean), in the following sequence: A + (B + C + D) fi A + B + (C + D) fi A + B + C + D. Assume that a species abcd¢ was widespread in A, B, C and D and speciated into four species, ...

Citation Context

...nder stationary, reversible and homogeneous conditions [for definitions, see e.g. Jayaswal et al. (2005)]. If these conditions are violated by data, then the risk of phylogenetic errors is increased (=-=Ho and Jermiin, 2004-=-; Jermiin et al., 2004). Alignments of nucleotides may vary compositionally in the sense that the composition may vary across sequences and/or across sites. In the first case, the sites would not have...

Citation Context

... molecular dating In total, we obtained 172 complete cytochrome b sequences from six Sus species. Haplotypes were found with CLEANUP software (Grillo et al., 1996) and phylogenetic analyses were then performed on 66 Sus sequences and one Phachocoerus africanus as outgroup. Of the 24 nucleotide substitution models tested by MRMODELTEST, the GTR+C gave the lowest score under the AIC criterion, and was thus used for the following Bayesian and maximum likelihood tests. Since it is well-known that variation in rates of evolution along different lineages can severely affect phylogenetic estimation (Ho and Jermiin, 2004), we inferred relationships among Sus species using the standard unrooted approach (every branch with its own rate of evolution), the strict molecular clock (the same rate for all branches), and the two relaxed clock phylogenetic models recently proposed by Drummond et al. (2006), in which rates of evolution are uncorrelated among branches but all belong to the same distribution (modeled as exponential or lognormal). We employed a pure birth model (birth rate being a hyper parameter with a uniform prior) and a uniform model (branch lengths with a uniform prior) to detect the possible effect of...

Citation Context

...xpected under a specified level of sampling error and are not truly comparable between data sets because their statistical power depends on factors such as the number of variable sites. Others (e.g., =-=Ho and Jermiin, 2004-=-) have articulated the need instead for simple methods for inferring whether any bias is in fact sufficient tosat Pennsylvania State U niversity on M ay 10, 2016 http://sysbio.oxfordjournals.org/ D ow...