International Journal of Marine Science, 2025, Vol.15, No.3, 130-143 http://www.aquapublisher.com/index.php/ijms 134 lineages. Comparing the results of nuclear DNA and mtDNA, if the topological structure is inconsistent, it often indicates a potential biological reason: it may be caused by historical hybridization causing the replacement of mtDNA, or it may be caused by the independent evolution of mitochondrial lineages. 3.3 Data processing and phylogenetic tree construction process Reconstructing phylogenetic relationships based on whole genome data usually requires multiple steps from raw data processing to phylogenetic tree inference. The first is sample collection and genome sequencing: select representative species and individuals for research questions, extract genomic DNA and perform high-throughput sequencing. The sequencing platform can be selected according to needs, such as Illumina short read length to obtain high coverage sequences, or PacBio/ONT long read length to facilitate crossing complex regions. After obtaining the original sequence, quality control is required to filter low-quality reads and adapter contamination .(Machado et al., 2022) For whole genome sequencing data, two strategies can be used: first, de novo assembly of the genome of each species, and then positioning and extracting conserved sequences on the genome for comparison; second, direct alignment to the reference genome without assembly or comparison through the k-mer matrix. In the study of the Spanish mackerel genus, since there is no perfect reference genome, the de novo assembly of mitochondria or specific genes has been used in the past. However, with the release of the reference genome of Spanish mackerel (such as S. guttatus), reads of other species can be aligned to the reference in the future, and homologous sites can be extracted from them for phylogeny. The next step is to identify and align homologous sequences. If there is a reference genome annotation, a list of single-copy orthologous genes shared by all species can be directly extracted, and then sequences can be extracted from each genome assembly or reads alignment for these genes. Subsequently, multiple sequence alignments are performed for each homologous gene or sequence, and regions with poor alignment quality are eliminated. In order to reduce computational pressure, supermatrices are often screened: for example, only single-copy genes or sites with rich variation information are retained. Then, phylogenetic analysis methods are selected according to the research purpose: splicing supermatrices to construct a "species tree" or a consensus method based on gene tree summary. The supermatrix method connects all gene alignments into long sequences and uses maximum likelihood (ML) or Bayesian methods to infer the overall phylogenetic tree. The advantage of the supermatrix method is that all data are used, and the disadvantage is that it may mask conflicts between different genes. Finally, the resulting phylogenetic tree is evaluated and corrected (Yáñez et al., 2023). It is necessary to pay attention to the support of each branch on the tree (such as bootstrap or posterior probability). Nodes with low support may require more data or different model testing. 4 Phylogenetic Reconstruction Results and Evolutionary Patterns 4.1 Optimization of classification structure and phylogenetic relationships With the help of whole genome data and multi-gene joint analysis, the phylogenetic relationship of the genus S. punctatus has been reconstructed more clearly recently. Overall, the new study supports the monophyletic origin of the genus S. punctatus and divides the internal species into several clades, optimizing the previous classification structure based on morphology or limited genes (Jeena et al., 2022). For example, multi-locus phylogenetic analysis roughly divides the species of this genus into two main evolutionary lineages: one includes the complex population of S. punctatus (S. guttatus and its cryptic relatives), S. multilineata, S. koreana, etc., and the other includes the Queensland S. punctatus and S. commerson, etc. Among them, the first lineage is further subdivided into multiple subbranches: the study found that the original broad Indo-Pacific horse mackerel (S. guttatus sensu lato) is actually composed of true horse mackerel (S. guttatus sensu stricto) and leopard horse mackerel (S. leopardus), and the two are clearly distinguished on the phylogenetic tree. True horse mackerel is mainly confined to the Bay of Bengal, which is consistent with the historical model origin, while leopard horse mackerel is widely distributed in other waters of the Indo-Pacific (Yang et al., 2023).
RkJQdWJsaXNoZXIy MjQ4ODYzNA==