2.4 De novo assembly and functional annotation

Adapter and low-quality reads were filtered by Fastp (version 0.19.7) (Chen et al., 2018) with default parameters in order to obtain clean data. Clean data was assembled to contigs using Trinity (v2.4.0) software with min_kmer_cov:3 and other default parameters (Grabherr et al., 2011). Based on Trinity splicing, transcripts were aggregated into clusters according to shared reads between transcripts by Corset (v4.6) (Davidson and Oshlack, 2014) with default parameters. Annotation to the NCBI Non-redundant Protein Sequences (NR), Nucleotide (NT), Eukaryotic Orthologous Groups (KOG), Swiss-Prot, Protein Families (PFAM), Gene Ontology (GO), and KEGG Orthology (KO) databases was carried out using BLASTX for all assembled unigenes (E-value≤1e-5) (Altschul et al., 1997; Moriya et al., 2007; Kanehisa et al., 2008; El-Gebali et al., 2019).