2011). To visualize genetic variability within and between morpho-species, analyses were performed with DnaSP v5.10 (Librado and Rozas 2009) estimated by pi (Nei 1987). Maximum likelihood and neigbor joining phylogenetic trees were inferred for each gene and for the concatenation of 28S rDNA, cox3, and tufA sequences LY2606368 clinical trial using the MEGA 5 software. Appropriate models of DNA substitution were detected with MEGA 5, using the three proposed statistics (AIC, AICc, and BIC). For most markers, the best-fit substitution model was the HKY model (Hasegawa et al. 1985), which distinguishes between transversion and transition rates with unequal base frequencies. For the tufA short
fragment data set, the K2P substitution model (Kimura 1980) was applied, differing from the HKY model by being based on equality of base frequencies, and for the rpl16 data set, the TN model (Tamura and Nei 1993) allowed different rates for two transitions (A-G and C-T) and constant rates for transversions with unequal base frequencies. Tree topologies were statistically tested by bootstrapping based on 1,000 replicates for both methods. The partial sequences of the nuclear 18S rDNA and the plastidial 16S rDNA and rbcL were strictly identical for all
strains of both morpho-species (Table 1), confirming that these genes are not suitable for discriminating between and within E. huxleyi and G. oceanica. Of the other genetic markers, the lowest value of nucleotide diversity (pi = 0.1 × 10−3) DAPT manufacturer was recorded with 28S rDNA sequences with a consistent 1 base pair differentiation between the two morpho-species. All other gene markers tested in this study exhibited
higher relative nucleotide substitution rates and polymorphisms, with the partial sequences of plastidial tufA (long; 6.4%) and mitochondrial dam (6.0%) exhibiting the highest degrees of variability for the set of cultures analysed (pi = 14.7 × 10−3 and 15.6 × 10−3, respectively; Table 1). While several of the markers tested exhibited sufficient variability to be potentially 4��8C suitable for barcoding and/or phylogenetic applications, full distinction of G. oceanica from E. huxleyi was not achieved with certain genes. The variability within tufA (long and short), petA, and cox1 (short) only partially delineated the two morpho-species, with interspecific overlap (i.e., polyphyly in phylogenetic reconstructions; Fig. 1). These markers exhibited a relatively high level of polymorphism (Table 1) highlighting microdiversity within each morpho-species. Previous studies using the plastid gene tufA also reported that microdiversity could be revealed within G. oceanica and E. huxleyi, but that these morpho-species cannot be clearly distinguished with this marker (Medlin et al. 2008, Cook et al. 2011). By contrast, consistent interspecific delineation was attained with the mitochondrial cox1 (long), cox2, cox3, rpl16, and dam markers. These mitochondrial markers also delineated consistent groups within E.