Genomic analyses reveal dead-end hybridization between two deeply divergent kiwifruit species rather than homoploid hybrid speciation.

Despite the importance of hybridization in evolution, the evolutionary consequence of homoploid hybridizations in plants remains poorly understood. Specially, homoploid hybridization events have been rarely documented due to a lack of genomic resources and methodological limitations. Actinidia zhejiangensis was suspected to have arisen from hybridization of A. eriantha and A. hemsleyana or A. rufa. However, this species was very rare in nature and exhibited sympatric distribution with its potential parent species, which implied it might be a spontaneous hybrid of ongoing homoploid hybridization. Here, we illustrate the dead-end homoploid hybridization and genomic basis of isolating barriers between A. eriantha and A. hemsleyana through whole-genome sequencing and population genomic analyses. Chromosome-scale genome assemblies of A. zhejiangensis and A. hemsleyana were generated. The chromosomes of A. zhejiangensis are confidently assigned to the two haplomes, each of them originates from A. eriantha and A. hemsleyana, respectively. Whole genome resequencing data reveals that A. zhejiangensis are mainly F hybrids of A. hemsleyana and A. eriantha and gene flow initiated about 0.98 million years ago, implying both strong genetic barriers and ongoing hybridization between these two deeply divergent kiwifruit species. Five inversions containing genes involved in pollen germination and pollen tube growth might account for the fertility breakdown of hybrids between A. hemsleyana and A. eriantha. Despite its distinct morphological traits and long recurrent hybrid origination, A. zhejiangensis does not initiate speciation. Collectively, our study presents new insights into homoploid hybridization in plants and provides genomic resources for evolutionary and functional genomic studies of kiwifruit.