分析流程总览
按执行顺序编号的分析流程。每个步骤目录保存该步 Gifu 实际运行的脚本，以及说明方法、输入、输出、参数与所用工具的 README.md；脚本保留原始文件名（部分脚本互相按名调用），步骤 README 指明该步的入口脚本。
| 阶段 | 步骤数 | 内容 |
|---|---|---|
01_assembly/ |
9 | 基因组组装、打磨、质控 |
02_annotation/ |
10 | 重复屏蔽、基因预测、过滤、基因分类、功能注释 |
03_telomere/ |
— | 端粒基序识别与末端验证 |
04_rDNA/ |
2 | rDNA 位点鉴定、阵列结构、拷贝数 |
05_centromere/ |
4 | 串联重复聚类与着丝粒界定 |
06_genome_comparison/ |
5 | 全基因组比对、结构变异、PAV、重排验证 |
07_transcriptome/ |
7 | bulk RNA-seq、组织图谱、WGCNA、Mfuzz、共表达 |
08_single_nucleus/ |
— | snRNA-seq 质控、整合、聚类、注释、虚拟敲除 |
阶段按编号顺序执行，01_assembly 内部顺序为 01_hifiasm → 02_purge_haplotigs → 03_remove_organelle → 04_telomere_check → 05_contig_ordering → 06_gapclose → 07_polish → 08_assembly_qc → 09_hic_scaffold；02_annotation 使用 01_assembly 产出的完成打磨的染色体级组装，03_telomere、04_rDNA、05_centromere 同样作用于完成组装的基因组且彼此独立，06_genome_comparison 需要两个材料都完成组装，07_transcriptome 与 08_single_nucleus 以注释为参考。
两个源码环境
Section titled “两个源码环境”脚本来自两位作者各自搭建的工作树，内部路径体现了这一点；工具版本见 environment/versions.tsv。
| 阶段 | 项目目录 | Conda 环境 |
|---|---|---|
01_assembly、02_annotation、03_telomere、04_rDNA、05_centromere |
{PROJ_04_LOTUS_GENOME}（注释脚本同时引用 {PROJ_LOTUS_ZC}） |
py3、geta-2.7.1、singularity 镜像 |
06_genome_comparison、07_transcriptome、08_single_nucleus |
{PROJ_LOTUS_HX} |
lotus、r_wgcna、singlecell |
| 脚本中的名称 | 指代 |
|---|---|
Gifu_v1.0.fasta |
Gifu T2T 组装（早期名称） |
Lotus_GifuT2T_v1.0.fasta |
同一套 Gifu T2T 组装（后期名称） |
Gifu_v1.0.fasta.mod.*、...mod.MAKER.masked |
由早期基名派生的 EDTA 输出 |
MG20_v1.0.fasta、Lotus_MG20T2T_v1.0.fasta |
MG20 T2T 组装，同样情况 |
同一套组装在项目不同时期被改名，并非不同的基因组，因此重复屏蔽、比对与注释脚本会混用这些名称，脚本可能用 Gifu_v1.0.fasta 建索引、同时读取 Lotus_GifuT2T 注释。ChiDou 与 ChiHei 是大豆品种名，不是 Lotus 材料：引用它们的脚本或属大豆分析（07_transcriptome 中的 Gm_* 步骤），或为 02_annotation/01_repeat_edta/LAI.sh 这类由大豆流程改写的脚本，用于 Gifu 前需替换基因组文件名。
脚本中没有机器相关的绝对路径，所有生产路径已替换为 {TOKEN} 占位符；全树做一次查找替换成自己的位置即可运行。占位符是纯文本，不是 shell 或 Python 语法：{TOKEN} 刻意不写作 ${TOKEN}``，以免被误认为已定义的变量。
| 占位符 | 含义 |
|---|---|
{PROJ_04_LOTUS_GENOME} |
基因组组装项目树（04_Lotus_genome） |
{PROJ_LOTUS_ZC} |
该作者更早的工作树（Project/Lotus） |
{PROJ_LOTUS_HX} |
RNA-seq / 单核项目树（Project/Lotus，第二位作者） |
{CONDA_PY3} |
conda 环境 py3 |
{CONDA_SINGLECELL}、{CONDA_LOTUS}、{CONDA_R_WGCNA} |
转录组与单核工作使用的三个 conda 环境 |
{CONDA_ZC}、{CONDA_HX} |
存放上述两个环境的两套 conda 安装 |
{SOFTWARE_ZC}、{SOFTWARE_HX}、{SOFTWARE_LOCAL} |
工具安装目录（BRAKER、HapHiC、StainedGlass 等） |
{HOME_ZC}、{HOME_HX}、{HOME_LOCAL} |
账号家目录 |
- 只有上表 15 个 token 是占位符。 脚本本身还含有
{sample}、{chrom}、{SP_NAME}、{WD}、{THREADS}、{PREFIX}、{label}、{n}等{...}表达式，属于作者自己的格式字符串变量或 shell heredoc 模板，在文件内定义与使用，不要替换；查找替换只替换上表列出的 token。 #CSUB、#SBATCH调度器指令、shebang 中的/usr/bin/env、计时封装中的/usr/bin/time、以及~形式的家目录路径（~/anaconda3/etc/profile.d/conda.sh、~/software/StainedGlass-0.6/）均保持原样，它们不是机器相关路径；tilde 形式可移植，同时假定~/software布局。
相关目录：environment/versions.tsv（工具版本、用途与来源）、third_party/（随仓库分发的第三方代码及其许可）、demo/（可运行小示例与期望输出）。
English
The analysis pipeline, numbered in execution order. Each step directory holds the scripts for that step **as they were run on the Gifu accession**, together with a `README.md` describing the method, inputs, outputs, parameters and the tools used; scripts keep their original filenames because several of them invoke one another by name, and the step README identifies the entry point for that step.Stage index
Section titled “Stage index”| Stage | Steps | What it covers |
|---|---|---|
01_assembly/ |
9 | genome assembly, polishing, quality control |
02_annotation/ |
10 | repeat masking, gene prediction, filtering, gene classification, functional annotation |
03_telomere/ |
— | telomere motif calling and terminal validation |
04_rDNA/ |
2 | rDNA locus identification, array structure, copy number |
05_centromere/ |
4 | tandem-repeat clustering and centromere delimitation |
06_genome_comparison/ |
5 | whole-genome alignment, structural variation, PAV, rearrangement validation |
07_transcriptome/ |
7 | bulk RNA-seq, tissue atlas, WGCNA, Mfuzz, co-expression |
08_single_nucleus/ |
— | snRNA-seq QC, integration, clustering, annotation, virtual knockout |
Execution order
Section titled “Execution order”Stages run in numeric order; inside 01_assembly the order is 01_hifiasm → 02_purge_haplotigs → 03_remove_organelle → 04_telomere_check → 05_contig_ordering → 06_gapclose → 07_polish → 08_assembly_qc → 09_hic_scaffold. 02_annotation consumes the polished, chromosome-level assembly produced by 01_assembly, 03_telomere, 04_rDNA and 05_centromere also operate on the finished assembly and can run independently of one another, 06_genome_comparison requires both accessions to be assembled, and 07_transcriptome and 08_single_nucleus use the annotation as their reference.
Two source environments
Section titled “Two source environments”The scripts come from the working trees of two co-authors who set up their analyses separately, and their internal paths reflect that; tool versions are listed in environment/versions.tsv.
| Stages | Project tree | Conda environments |
|---|---|---|
01_assembly, 02_annotation, 03_telomere, 04_rDNA, 05_centromere |
{PROJ_04_LOTUS_GENOME} (annotation scripts also reference {PROJ_LOTUS_ZC}) |
py3, geta-2.7.1, singularity images |
06_genome_comparison, 07_transcriptome, 08_single_nucleus |
{PROJ_LOTUS_HX} |
lotus, r_wgcna, singlecell |
Genome naming
Section titled “Genome naming”| Name seen in the scripts | Refers to |
|---|---|
Gifu_v1.0.fasta |
the Gifu T2T assembly (earlier name) |
Lotus_GifuT2T_v1.0.fasta |
the same Gifu T2T assembly (later name) |
Gifu_v1.0.fasta.mod.*, ...mod.MAKER.masked |
EDTA outputs derived from the earlier basename |
MG20_v1.0.fasta, Lotus_MG20T2T_v1.0.fasta |
the MG20 T2T assembly, same situation |
The same assemblies were renamed at different periods of the project and are not different genomes, so repeat-masking, alignment and annotation scripts mix these names freely, and a script may build an index from Gifu_v1.0.fasta while reading a Lotus_GifuT2T annotation. ChiDou and ChiHei are soybean cultivar names, not Lotus accessions: scripts that reference them are either soybean analyses (the Gm_* steps in 07_transcriptome) or, as with 02_annotation/01_repeat_edta/LAI.sh, adapted from a soybean run with the genome filename substituted before use on Gifu.
Path placeholders
Section titled “Path placeholders”No script contains a machine-specific absolute path: every production path has been replaced by a {TOKEN} placeholder, and one search-and-replace over the whole tree substitutes your own locations. Placeholders are plain text, not shell or Python syntax: {TOKEN} is deliberately not ${TOKEN}``, so a bare token cannot be mistaken for a variable that is already defined.
| Token | Stands for |
|---|---|
{PROJ_04_LOTUS_GENOME} |
the genome-assembly project tree (04_Lotus_genome) |
{PROJ_LOTUS_ZC} |
the earlier working tree of the same author (Project/Lotus) |
{PROJ_LOTUS_HX} |
the RNA-seq / single-nucleus project tree (Project/Lotus, second author) |
{CONDA_PY3} |
conda environment py3 |
{CONDA_SINGLECELL}, {CONDA_LOTUS}, {CONDA_R_WGCNA} |
the three conda environments used for the transcriptome and single-nucleus work |
{CONDA_ZC}, {CONDA_HX} |
the two conda installations that hold those environments |
{SOFTWARE_ZC}, {SOFTWARE_HX}, {SOFTWARE_LOCAL} |
tool installation directories (BRAKER, HapHiC, StainedGlass, …) |
{HOME_ZC}, {HOME_HX}, {HOME_LOCAL} |
the account home directories |
- Only the 15 tokens in the table above are placeholders. Several scripts already contained
{...}expressions of their own —{sample},{chrom},{SP_NAME},{WD},{THREADS},{PREFIX},{label},{n}and others. Those are the authors’ format-string variables or shell heredoc templates, defined and used inside the file; do not substitute them — a search-and-replace replaces only the tokens listed above. #CSUB/#SBATCHscheduler directives,/usr/bin/envin shebangs,/usr/bin/timein the timing wrappers, and home-relative paths written with~(~/anaconda3/etc/profile.d/conda.sh,~/software/StainedGlass-0.6/) were all left untouched: they are not machine-specific; the tilde form is portable and assumes a~/softwarelayout.
Related directories
Section titled “Related directories”Related directories: environment/versions.tsv (tool versions, purposes and sources), third_party/ (third-party code redistributed with this repository, and its licences), demo/ (small runnable examples with expected output).
| 步骤 | 内容 |
|---|---|
| 01 组装 | — |
| 02 注释 | — |
| 03 端粒 | — |
| 04 rDNA | — |
| 05 着丝粒 | — |
| 06 比较基因组 | — |
| 07 转录组 | — |
| 08 单核转录组 | — |