In this entry, researchers used Treemix to confirm the gene flow from M. amblycephala to M. skolkovii. Using M. hoffmanni as the outgroup, among the four Megalobrama species, only one migration event was inferred from the M. amblycephala to the M. skolkovii with approximately 50.77% DNA gene flow from M. amblycephala to M. skolkovii (Figure 3A). Moreover, the migration patterns of these two species’ geographic populations demonstrated that LZL, DTL, TEL, YNL, and PYL M. amblycephala populations introgressed into the FY, SG, QT, and JS M. skolkovii populations. Notably, the migration weight of M. amblycephala introgressed into QT and JS populations was greater than that of the FY and SG populations.

The genetic diversity (π) of M. skolkovii and M. hoffmanni were estimated to be 3.827 × 10⁻³ and 3.324 × 10⁻³, respectively, which was relatively high in comparison to M. amblycephala (2.072 × 10⁻³) and M. pellegrini (1.888 × 10⁻³). The low genetic differentiation (Fst) between M. skolkovii and M. pellegrini (0.148), and the high Fst between M. hoffmanni and M. amblycephala (0.4091), M. skolkovii (0.3516), and M. pellegrini (0.4259) (Figure 3B) are consistent with the phylogeny analysis. Between the four Megalobrama species, linkage disequilibrium (LD) and correlation coefficient (r²) values were calculated. The decay of LD reached half the maximum average r² at a distance of 24 Kb, 4 Kb, 1.8 Kb, and 0.2 Kb for M. amblycephala, M. pellegrini, M. skolkovii, and M. hoffmanni, respectively (Figure 3C). Therefore, M. skolkovii and M. hoffmanni displayed a faster LD decay rate than M. amblycephala and M. pellegrini.

The split times based on the relative cross-coalescent rates (RCCR) among the four Megalobrama species reached 0.5 suggesting a split between M. hoffmanni and other species 3–5 Mya. The cross-coalescence analysis suggested a decline to 0.5 between M. amblycephala and M. skolkovii or M. pellegrini at ~1.3 Mya, and a decline to 0.5 between M. skolkovii and M. pellegrini around 0.3 - 0.4 Mya (Figure 4A). The PSMC method used to reconstruct the demographic history of six Megalobrama subgroups indicated that the effective population size (Ne) peak of M. amblycephala and M. pellegrini was 2.5 Mya and 4 Mya, respectively, compared to 3 - 4 Mya of the Ne peak for the M. skolkovii subgroups. After that, the FY M. skolkovii subgroup continued to shrink, whilst the Ne curves of other M. skolkovii subgroup split at 0.2 Mya expanded (Figure 4B). Moreover, the two Ne peaks of M. hoffmanni occurred at 0.3 Mya and 5 Mya, respectively, during the middle pleistocene geological period (Figure 4C). Interestingly, the split of Ne curves of the two M. hoffmanni subgroups occurred about 0.4 Mya, indicating a population divergence at this time. The ancestral reconstruction analysis revealed two different biogeographic evolutionary processes to investigate the ancestral distribution of genus Megalobrama. According to the BBM analysis, the genus Megalobrama was originally distributed in the Pearl River and then spread to Hainan Island and Northern China. However, the analysis based on the S-DIVA and DEC models showed that the Megalobrama ancestors originally inhabited the Pearl River and Yangtze River, before spreading to the Amur and Wusuli Rivers.

To investigate the potential selective signals during M. amblycephala dietary adaptation, researchers scanned the genomic regions based on genome-wide calculations for selective sweeps by estimating Fst, PiR, and XP-EHH values. Candidate genes were discovered in the common region of top 5% Fst, High/Low top 5% PiR, and High/Low top 5% XP-EHH, inferred from the comparisons between M. amblycephala and the other three species (Figures 5A). Fatty acid degradation, glycerolipid metabolism, beta-alanine metabolism, arginine and proline metabolism, histidine metabolism, insulin secretion, and lysine degradation, among other metabolic processes were associated with a significant portion of the candidate genes, according to GO and KEGG analyses. M. amblycephala mainly feeds on high-fiber and low-energy aquatic vascular plants. The candidate genes of M. amblycephala were enriched in fatty acid degradation and corresponding upstream and downstream pathways. For instance, Aldh3a2 and Acss3 genes in fatty acid α-oxidation, Hadhb gene in fatty acid β-oxidation, Akt2 gene in insulin signaling pathway, Aldh3a2 gene in glycolysis/gluconeogenesis, Gbe1 gene in starch and sucrose metabolism, Acsbg2 gene in fatty acid biosynthesis, and Aldh3a2 gene in valine, leucine, and isoleucine degradation (Figure 6A）. Moreover, these genes exhibited a high expression pattern in the liver and/or spleen tissue of M. amblycephala compared with M. hoffmanni (Figure 6B-E,b-e).

Among the three omnivorous Megalobrama species, the food composition of M. hoffmanni is rich in zoobenthos. The common regions of top 5% Fst, High top 5% PiR, and Low top 5% XP-EHH were the selected regions inferred from a comparison of M. hoffmanni and M. amblycephala, which covered a total of 75 selected regions with a length of 4.14 Mb and including 126 genes in M. hoffmanni. These genes were found to be abundant in cellular and metabolic processes including taste transduction, fatty acid elongation, biosynthesis of unsaturated fatty acids, and biosynthesis of amino acids according to GO and KEGG enrichment analysis. Interestingly, researchers found some candidate genes (Cyp46a1 and Baat) involved in cholesterol metabolism (Figures 5B, 6A). The results indicated that compared with M. amblycephala, the Cyp46a1 gene was highly expressed in the liver of M. hoffmanni, but not in the spleen (Figure 6F,f). Furthermore, functional annotation indicated that the umami taste receptor gene Tas1r1 was abundant in the M. hoffmanni sensory system.