Genomic annotation of the yellow drum

Structural and functional gene annotation of the
assembled genome was predicted using de novo, homolog-based, and PacBio
full-length transcriptome-based strategies. The repetitive sequences detected
above were masked before de novo prediction annotation. We used both Augustus
(augustus-3.2.1) 26 and Genscan 27 to formulate de novo predictions, yielding
38,857 and 38,807 protein-coding genes, respectively. The protein sequences of
closely related fish species, including D. rerio, G. morhua, G.
aculeatus, O. niloticus, O. latipes, T. rubripes, T.
nigroviridis, and X. maculatus, were downloaded from Ensembl 28 to facilitate
homolog-based gene prediction. All of the protein sequences were mapped to the
yellow drum genome using the Basic Local Alignment Search Tool All (BLASTALL). We
predicted 23,119 genes based on the PacBio full-length transcriptomic data,
which were then used in transcriptome-based annotation. We then integrated the
results of the de novo predictions, homolog-based annotations, and
transcriptome-based annotations using GLEAN 29, finally obtaining a dataset containing
27,069 protein-coding genes