作用：处理 Gifu PacBio Iso-Seq 文库：isoseq3 refine 去引物、isoseq3 cluster 得到高质量全长一致序列，再比对到 T2T 组装得到长读转录证据 BAM。
入口：isoseq.sh。
输入：脚本内硬编码的 sample1 = {PROJ_LOTUS_ZC}`/data/Gifu/isoseq/hifi_reads/P22TR251406648-1-bc12_r84069_20250617_081306_1_A01.hifi_reads.bam` 与同前缀的 `primer1`（`.primer.fasta`）；参考基因组 {PROJ_LOTUS_ZC}/Gifu/00.assemble_result/Gifu_v1.0.fasta。
输出：Gifu.refined.bam、Gifu.clustered.bam、Gifu.clustered.fastq、Gifu_isoforms.sam、Gifu_isoforms.sorted.bam（含索引），后者被 04_braker/Gifu_braker.sh 使用。
运行：bash isoseq.sh；仅 bc12 文库生效，覆盖样本 1–4 的 bamtools merge 循环与 TAMA/gffread/TransDecoder 段均被注释。
工具：isoseq3 refine "$sample1" "$primer1" Gifu.refined.bam、cluster Gifu.refined.bam Gifu.clustered.bam --verbose --use-qvs；minimap2 -ax splice -uf -k14 <genome> Gifu.clustered.fastq；samtools fastq、sort -o Gifu_isoforms.sorted.bam、index。
English
**Purpose**: Process the Gifu PacBio Iso-Seq library — `isoseq3 refine` removes primers, `isoseq3 cluster` builds high-quality full-length consensus isoforms, and the isoforms are aligned to the T2T assembly to give a long-read transcript-evidence BAM.
**Entry point**: `isoseq.sh`.
**Inputs**: the hardcoded `sample1` = ``{PROJ_LOTUS_ZC}`/data/Gifu/isoseq/hifi_reads/P22TR251406648-1-bc12_r84069_20250617_081306_1_A01.hifi_reads.bam` and `primer1` = the `.primer.fasta` with the same prefix; reference genome ``{PROJ_LOTUS_ZC}`/Gifu/00.assemble_result/Gifu_v1.0.fasta`.
**Outputs**: `Gifu.refined.bam`, `Gifu.clustered.bam`, `Gifu.clustered.fastq`, `Gifu_isoforms.sam`, `Gifu_isoforms.sorted.bam` (plus index), the last consumed by `04_braker/Gifu_braker.sh`.
**Run**: `bash isoseq.sh`; only the bc12 library is used, while the loop over samples 1–4 with `bamtools merge` and the TAMA/`gffread`/TransDecoder blocks are commented out.
**Tools**: isoseq3 `refine "$sample1" "$primer1" Gifu.refined.bam`, `cluster Gifu.refined.bam Gifu.clustered.bam --verbose --use-qvs`; minimap2 `-ax splice -uf -k14 Gifu.clustered.fastq`; samtools `fastq`, `sort -o Gifu_isoforms.sorted.bam`, `index`.