3.2 Genetic structure and differentiation
The molecular evolution model with the gamma shape parameter (HKY+I+G) was found to be the best substitution model for the COI and Cytb gene datasets by the jModelTest. From this result, a BI tree was built to determine phylogenetic relationship across populations (Figure S1). The BI tree was dispersed with haplotypes from each population and lacked well-supported groups. Neither significant genealogy branches nor haplotype clusters could be identified in relation to the sampling locations.
Network analysis and the phylogenetic tree showed similar results. The connection between several haplotypes resembled a star, with certain prominent haplotypes like Hap_1, Hap_2, Hap_4, and Hap_22 (Figure 2). Unnoticeable clades in the network diagram of reticulations do not correlate with sampling locations, suggesting a substantial gene flow among populations and recent population expansion. It was found that relationships between populations were not linked to geological networks, but instead to the haplotypes that were present in each population. According to these results, there was no obvious phylogeographical pattern of L. spadiceus in the South China Sea.
The degree of genetic variation between populations was evaluated usingF st pairwise comparisons. TheF st values were typically low and even negative, as shown in Table 2, only the F st values between LZ population and other populations were higher and significant (P < 0.05). The genetic distance between and within populations was at the same level, with little differentiation (Table 2). The range under investigation had no significant genetic structure. Hierarchical AMOVA analysis revealed that genetic variation in all populations existed within populations, while only a small proportion of genetic variation could be attributed to differences between populations (Table 3).