Cephus runcator Konow, 1896

Budak, Mahir, Korkmaz, E. Mahir & Basibuyuk, Hasan H., 2011, A molecular phylogeny of the Cephinae (Hymenoptera, Cephidae) based on mtDNA COI gene: a test of traditional classification, ZooKeys 130, pp. 363-378 : 366-368

Cephus runcator Konow, 1896

Material examined.

Turkey: Edirne [40°39.32'N, 26°17.82'E] 50 m, 07.05.2010, 6♀, 1♂.

Distribution.

Turkey, S. E. Europe.

The complete alignment of the partial mitochondrial COI gene sequences from 68 cephid specimens, including representatives of these two new records, resulted in a fragment containing 658 base pairs, among which 287 nucleotide positions were variable and 223 sites of which were parsimony-informative. The analyzed sequences correspond to a functional mitochondrial gene region because of the presence of singular peaks in each chromatograph and absence of in–del and premature stop codons, and presence of the highest nucleotide substitutions at the third codon position ( Avise 1994). The percentages of nucleotide composition at each codon position are variable (Fig. 1). The mean frequency of COI sequences used in the analyses showed a bias of A + T (T 37.0%, C 15.2%, A 33.9% and G 16.0%), which is similar to other reported members of Hymenoptera ( Jermiin and Crozier 1994, Dowton and Austin 1995, 1997, Leys et al. 2000, Danforth et al. 2003). The A + T content at the third, second and first codon positions are 90.7%, 59.8%, and 61.9%, respectively. The nucleotide G has lowest (1.0%) and the A highest content (52.8%) at the third codon positions. The distribution of polymorphic sites for all cephid species shows that the majority of substitutions are at synonymous sites. The vast majority of synonymous substitutions are also found at third codon positions with a rate of 87.88% for the Cephinae . The first and the second positions are relatively more conserved in comparison with the third position.

The numbers of base substitutions per site from averaging over all sequence pairs between genera are shown in Table 2. The least diverged genera appears to be Characopygus and Pachycephus (p= 0.062) and, the most are Hartigia and Syristra (p= 0.161) also with highest standard error value of 0.017.

All three codon positions in the partial COI gene were analyzed for saturation, achieved by plotting the number of observed substitutions against the model TN93 genetic distance estimates. The scattergrams (Fig. 2 a–c) showed that transitions and transversions for the first, second and third codons of the partial COI gene increased with the genetic distance, but considerable scattering was also observed. In addition, a similar plot of the third codon transition of the COI gene (Fig. 2d) suggested that saturation of transition occurred between certain pairs of the taxa, which may lead to higher levels of homoplasy ( Kumar et al. 2001; Zhang et al. 2008).

Result of likelihood mapping is presented in Fig. 3. High dichotomic phylogenetic signal was detected in the dataset. The percentage of the quartets suggesting a star- or network- like phylogeny is 9.9%, indicating that data are reliable for a dichotomic phylogenetic analysis ( Schmidt 2009). For ML analysis GTR+I+G models showed a significantly better fit than the other less complicated models for the COI dataset. Maximum likelihood analyses under the same model of evolution resulted in topologies with lnL = - 5570.6831 in RAxML, which were very close to the BI tree. Bayesian inference under the GTR+I+G model resulted in a topology with mean lnL = - 5347.963. Posterior probability values from the BI were congruent with ML bootstrap support. ML and BI analyses generated a tree with almost the same overall topology (Fig. 4). Equally weighted parsimony analysis of the 287 parsimony-informative characters produced 12 most parsimonious trees with a length of 1065 steps (Homoplasy Index = 0.608, Retention Index = 0.392 and Consistency Index = 0.392). These equally parsimonious solutions were due to differences in terminal branches. However, the branching pattern of bootstrap tree was comb-like and recovered almost no original branches. Considering that this may be due to many synonymous changes in the third codon position, we run an analysis excluding the third codon position from the data. The analysis produced 60 equally parsimonious trees with a length of 200 and the bootstrap application was also resulted with no support to branching pattern of original trees. This may be partly attributed to the nature of data and relatively a short sequence not sufficient to detect phylogenetic signal under parsimony interference. Therefore, we do not present any MP trees here.