Unlike birds and mammals, many teleosts have homomorphic sex chromosomes, and changes in the chromosome carrying the sex-determining locus, termed "turnovers", are common. Recent turnovers allow studies of several interesting questions. One question is whether the new sex-determining regions evolve to become completely non-recombining, and if so, how and why. Another is whether (as predicted) evolutionary changes that benefit one sex accumulate in the newly sex-linked region. To study these questions, we analyzed the genome sequences of two seahorse species of the Syngnathidae, a fish group in which many species evolved a unique structure, the male brood pouch. We find that both seahorse species have XY sex chromosome systems, but their sex chromosome pairs are not homologs, implying that at least one turnover event has occurred. The Y-linked regions occupy 63.9% and 95.1% of the entire sex chromosome of the two species and do not exhibit extensive sequence divergence with their X-linked homologs. We find evidence for occasional recombination between the extant sex chromosomes that may account for their homomorphism. We argue that these Y-linked regions did not evolve by recombination suppression after the turnover, but by the ancestral nature of the low crossover rates in these chromosome regions. With such an ancestral crossover landscape, a turnover can instantly create an extensive Y-linked region. Finally, we test for adaptive evolution of male pouch-related genes after they became Y-linked in the seahorse.

Seahorses belong to the teleost family Syngnathidae that evolved a distinct body plan and unique male pregnancy compared to other teleosts. As a classic model for studying evolution of viviparity and sexual selection of teleosts, seahorse species still lack a publicly available high-quality reference genome. Here, we generated the genome assembly of the big-belly seahorse, Hippocampus abdominalis with long-read and Hi-C technologies. We managed to place over 99% of the total length of 444.7 Mb of assembled genome into 21 linkage groups with almost no gaps. We reconstructed a phylogenomic tree with the big-belly seahorse genome and other representative Syngnathidae and teleost species. We also reconstructed the historical population dynamics of four representative Syngnathidae species. We found the gene families that underwent expansion or contraction in the Syngnathidae ancestor were enriched for immune-related or ion transporter gene ontology terms. Many of these genes were also reported to show a dynamic expression pattern during the pregnancy stages of H. abdominalis. We also identified putative positively selected genes in the Syngnathidae ancestor or in H. abdominalis, whose mouse mutants are enriched for abnormal craniofacial and limb morphological phenotypes. Overall, our study provides an important genome resource for evolutionary and developmental studies of seahorse species, and candidate genes for future experimental works.