In 2024, we reported the first chromosome-level genome for Euglena gracilis. Later in 2025, Nomura et al. reported the noncanonical splicing code and generated an improved annotation for Euglena gracilis. Inspired by their work, we reported an improved annotation based on saturated Iso-seq backbone, named as "EugrW". To improve the annotation, we first generated 23 M Iso-seq reads from RNA samples across various growth conditions of Euglena gracilis. After using the standard Iso-Seq workflow with SMRT Link v25.3 for primer removal, demultiplexing, refine, clustering and mapping, the mapped reads were collapsed using tama_collapse with the following parameters: “-a 10 -m 10 -z 10 -x no_cap -e longest_ends”. Then, these gene models were enhanced by incorporating 6 RNA-seq datasets (SRR3195332, SRR3195334, SRR3195335, SRR3195338, SRR3195339, SRR3195340) using tama_merge with default parameters. Genes without Iso-seq support and only supported by one RNA-seq dataset were excluded. Open reading frames were predicted using TransDecoder-v5.7.1 with the parameters: “--single_best_only --retain_pfam_hits --retain_blastp_hits”, while the ORFs were searched against the Pfam-A database (release 38.1) using HMMER-v3.4 and the OthroDB v12.2 (Euglenozoa and Chlorophyta clades) datasets using BLASTP-v2.17.0 with the parameter: “-evalue 1e-5”.