2.5 Genome assembly
The error of the nanopore clean reads was first corrected using NextDenovo (https://github.com/ Nextomics/NextDenovo), with the seed cutoff set at 22 k. The C. undulatus genome was assembled using SMARTdenovo (https://github.com/ruanjue/smartdenovo). The nanopore- assembled genome was polished in two runs for error-corrected long reads using the Illumina DNA short reads by NextPolish (Hu et al. 2020). The genes in the assembled genome were predicted using BUSCO (Simão et al. 2015) with the vertebrata_odb9 database. The integrity of the genome was assessed using the Illumina short reads by BWA (Li & Durbin 2010) and the Illumina RNA-seq reads of multiple tissues by hisat2 (Kim et al.2015) aligned against the assembled genome. The accuracy rate of single base was validated by SNP calling using FreeBayes (Garrison & Marth 2012).