作用：把 WGCNA brown 与 lightcyan1 模块的基因沿结瘤发育序列（root_uni → root_hpi48 → nodule_dpi10 → dpi21 → dpi45）做软聚类，再把同一批簇投影到根序列上。
入口：Mfuzz.R
输入：工作目录中的 Gifu_all_samples.tpm.tsv、MG20_all_samples.tpm.tsv，以及 ./result/06.modules_expr/module_brown.csv 与 module_lightcyan1.csv。
输出：result/12.Mfuzz_out/ 下的 Dmin_plot.pdf、Mfuzz_stability.pdf、Mfuzz_test.pdf、cluster1.genes.list…cluster4.genes.list、Mfuzz_membership.tsv、Mfuzz_plot.pdf、Mfuzz_plot2.pdf、Mfuzz_root_projection_plot.pdf、Mfuzz_root_projection_plot2.pdf。
运行：Rscript Mfuzz.R（无参数；工作目录须同时持有两张 TPM 矩阵与 WGCNA.R 生成的 result/ 树）。
工具：R 4.4.3 + Mfuzz（mestimate、Dmin、mfuzz、mfuzz.plot、mfuzz.plot2）、Biobase、tidyverse；参数：阶段均值 TPM 取 log1p、filter.NA(thres = 0.25)、fill.NA(mode = "mean")、filter.std(min.std = 0)、Dmin(crange = seq(4, 20), repeats = 3)、最终 c = 4（set.seed(1234)）。
English
**Purpose**: Soft-cluster the genes of the WGCNA brown and lightcyan1 modules along the nodule developmental series (root_uni → root_hpi48 → nodule_dpi10 → dpi21 → dpi45) and project the same clusters onto the root series.
**Entry point**: `Mfuzz.R`
**Inputs**: `Gifu_all_samples.tpm.tsv` and `MG20_all_samples.tpm.tsv` in the working directory, plus `./result/06.modules_expr/module_brown.csv` and `module_lightcyan1.csv`.
**Outputs**: under `result/12.Mfuzz_out/` — `Dmin_plot.pdf`, `Mfuzz_stability.pdf`, `Mfuzz_test.pdf`, `cluster1.genes.list` … `cluster4.genes.list`, `Mfuzz_membership.tsv`, `Mfuzz_plot.pdf`, `Mfuzz_plot2.pdf`, `Mfuzz_root_projection_plot.pdf` and `Mfuzz_root_projection_plot2.pdf`.
**Run**: `Rscript Mfuzz.R` (no arguments; the working directory must hold both TPM matrices and the `result/` tree produced by `WGCNA.R`).
**Tools**: R 4.4.3 with Mfuzz (`mestimate`, `Dmin`, `mfuzz`, `mfuzz.plot`, `mfuzz.plot2`), Biobase and tidyverse; settings as coded: `log1p` on stage-mean TPM, `filter.NA(thres = 0.25)`, `fill.NA(mode = "mean")`, `filter.std(min.std = 0)`, `Dmin(crange = seq(4, 20), repeats = 3)`, and the final clustering `c = 4` with `set.seed(1234)`.