Pan-genome analysis of three main Chinese chestnut varieties

Abstract

Chinese chestnut (Castanea mollissima Blume) is one of the earliest domesticated plants of high nutritional and ecological value, yet mechanisms of C. mollissima underlying its growth and development are poorly understood. Although individual chestnut species differ greatly, the molecular basis of the formation of their characteristic traits remains unknown. Though the draft genomes of chestnut have been previously released, the pan-genome of different variety needs to be studied. We report the genome sequence of three cultivated varieties of chestnut herein, namely Hei-Shan-Zhai-7 (H7, drought-resistant variety), Yan-Hong (YH, easy-pruning variety), and Yan-Shan-Zao-Sheng (ZS, early-maturing variety), to expedite convenience and efficiency in its genetics-based breeding. We obtained three chromosome-level chestnut genome assemblies through a combination of Oxford Nanopore technology, Illumina HiSeq X, and Hi-C mapping. The final genome assemblies are 671.99 Mb (YH), 790.99 Mb (ZS), and 678.90 Mb (H7), across 12 chromosomes, with scaffold N50 sizes of 50.50 Mb (YH), 65.05 Mb (ZS), and 52.16 Mb (H7). Through the identification of homologous genes and the cluster analysis of gene families, we found that H7, YH and ZS had 159, 131, and 91 unique gene families, respectively, and there were 13,248 single-copy direct homologous genes in the three chestnut varieties. For the convenience of research, the chestnut genome database1 was constructed. Based on the results of gene family identification, the presence/absence variations (PAVs) information of the three sample genes was calculated, and a total of 2,364, 2,232, and 1,475 unique genes were identified in H7, YH and ZS, respectively. Our results suggest that the GBSS II-b gene family underwent expansion in chestnut (relative to nearest source species). Overall, we developed high-quality and well-annotated genome sequences of three C. mollissima varieties, which will facilitate clarifying the molecular mechanisms underlying important traits, and shortening the breeding process.

Chromosome-level genome assembly provides new insights into Japanese chestnut (Castanea crenata) genomes

“…A total of 421.37 Mb of the sequence was annotated as repetitive sequences in the genome assembly of cultivar 'Tsukuba', which was similar to the genome assembly of cultivar 'Ginyose' (Shirasawa et al, 2021). The two Japanese chestnut genome assemblies exhibit a repetitive rate similar to that of the genome assembly of Chinese chestnut (431.41 Mb,55.74%;437.75 Mb,64.38%;423.16 Mb,53.49%;442.76 Mb,64.43%) described by Sun et al (Sun et al, 2020) and Hu et al (Hu et al, 2022), but slightly different from that found in research by Wang et al (366.84 Mb,53.24%) (Wang et al, 2020b) and Xing et al (390 Mb,49.69%) (Xing et al, 2019). This difference might be caused by the exclusion of duplicated contigs in our research or the use of different annotation methods.…”

“…However, 69,980 high-confidence genes were predicted in the genome of the cultivar 'Ginyose' (Shirasawa et al, 2021). Compared with the former genome annotation of Chinese chestnut (Table 2), our genome annotation of Japanese chestnut predicted a similar number of gene models as the research conducted by Sun et al (Sun et al, 2020) and predicted more gene models than the research performed by Wang et al (Wang et al, 2020b), Xing et al (Xing et al, 2019) and Hu et al (Hu et al, 2022). Compare the BUSCO scores of these predicted gene sets, more than 193 complete BUSCOs were predicted in the Genome of 'Tsukuba' than that in the Genome of 'Ginyose' (Table 2).…”

“…Three methods were used to assess the quality of the primary assembly. First, the Illumina short reads were mapped back to the primary assembly, and 96.91% (521.82 million reads out of 538.47 million reads) of the total reads were properly paired in the primary assembly, which is significantly higher than several Chinese chestnut varieties (84.24% ('Yan-Hong'), 90.36% ('Yan-Shan-Zao-Sheng'), and 89.98% ('Hei-Shan-Zhai-7')) (Hu et al, 2022). Second, Benchmarking Universal Single-Copy Orthologs (BUSCOs) analysis showed that 97.8% of the complete BUSCOs were generated in the primary assembly, even though 17.3% of these complete BUSCOs were duplicated (Table 1) (Seppey et al, 2019).…”

“…( Xing et al., 2019 ) and Hu et al. ( Hu et al., 2022 ). Compare the BUSCO scores of these predicted gene sets, more than 193 complete BUSCOs were predicted in the Genome of ‘Tsukuba’ than that in the Genome of ‘Ginyose’ ( Table 2 ).…”

“…Marker-assisted breeding and genomic selection can speed up the breeding cycle, but these methods require one or several high-quality reference genomes. Although seven Chinese chestnut genome assemblies are available publicly ( Xing et al., 2019 ; Staton et al., 2020 ; Sun et al., 2020 ; Wang et al., 2020b ; Hu et al., 2022 ), only two Japanese chestnut genome assemblies have recently been released by NCBI (GCA_019972055.1 and GCA_020976635.1); further research is still needed on the genome variations in Japanese chestnut. One of the released Japanese chestnut genomes is for cultivar ‘Ginyose’ and was generated at the scaffold to chromosome-level ( Shirasawa et al., 2021 ); the other is a draft genome assembly that was generated from one 400-years-old native tree in South Korea.…”