Genome assembly

We analyzed the genome size of the sequenced individuals by flow cytometry (BD FACSCalibur, BD Bioscience, USA) using tomato and maize as internal controls⁵⁰. The CSS HD genome was assembled by incorporating sequencing data from PacBio circular consensus sequencing technology⁵¹ and the Hi-C method. First, HiFi reads were assembled with Hifiasm with the default parameters. We assembled two levels of chromosome-scale genomes, including a monoploid genome and an allele-defined haplotype-resolved assembly. Briefly, for monoploid assembly, although Purge_dups was contained in Hifiasm, we checked the read depth and filtered the primary contigs of the initial Hifiasm assembly by Purge_dups (v1.25) with the default parameters and evaluated the results by the Benchmarking Universal Single-Copy Orthologs (BUSCO) completeness²⁷ and duplication score. Hi-C reads were assessed by the HiC-Pro (v2.11.4) program⁵² and were uniquely mapped to contig assemblies. Meanwhile, Juicer tools (v1.11.08)⁵³ and 3D-DNA pipelines (v180114)⁵⁴ were used to detect and correct misassembled contigs. To distribute contigs into the appropriate groups, we aligned the set of Hi-C corrected contigs against the “Shuchazao” chromosome-scale assembly using RaGOO (v1.1). Finally, the ALLHiC optimize algorithm (v0.9.13) was used to adjust the order and orientation of contigs in each group. We merged the haplotig sequence and alternative assembly to redo Purge_dups (v1.25) with the cutoff parameter “2 6 10 12 20 70”. The purged primary contigs and alternative contigs were merged and regarded as a draft contig assembly for the monoploid genome. The resulting contigs were subjected to haplotype phasing using the ALLHiC algorithm with default parameters, and the monoploid genome sequences were selected as a reference to identify allelic contigs. Chromosome-level haplotype A and haplotype B of CSS HD were fully resolved and released. In addition, we calculated the heterozygosity through GenomeScope2⁵⁵ with 33 k-mers. Chromosome localization and collinearity analysis of genes were visualized in TBtools⁵⁶.