作用：把 SNF 家族基因与候选基因列表在两种材料的根、接种根与结瘤样本中做相关，保留通过 Bonferroni 校正 p 值与相关系数阈值的基因对并写出网络。
入口：co-exp_psych.R
输入：参数 go_id_list.txt（首列为候选基因 ID）与 out_dir（输出目录）；../../../Gifu_all_samples.tpm.tsv、../../../MG20_all_samples.tpm.tsv；工作目录中的 all_new_genes.list、SNF_gene.list 与同源表 All_species_merged_by_GifuT2T.tsv。
输出：<out_dir>/SNF_vs_ALL/ 与 <out_dir>/SNF_vs_New/，各含 edges_bonferroni.txt、nodes.txt、network_igraph.rds；<out_dir>/network_gene_composition.txt 给出各节点类别的计数。
运行：Rscript co-exp_psych.R go_id_list.txt output_dir（须从两张 TPM 矩阵之上三层的目录运行）。
工具：R 4.4.3 + psych（corr.test(method = "pearson", adjust = "none")）、igraph、data.table 与 tidyverse；参数：样本列匹配 root_hpi48/root_uni/nodule，最小 TPM > 0.5，alpha = 0.05、p_cutoff = alpha / nrow(edges_all)，保留 p < p_cutoff 且 cor > 0.7 的边。
English
**Purpose**: Correlate SNF-family genes with a candidate gene list across the root, inoculated-root and nodule samples of both accessions, keep pairs passing a Bonferroni-corrected p-value and a correlation cutoff, and write the resulting networks.
**Entry point**: `co-exp_psych.R`
**Inputs**: the arguments `go_id_list.txt` (first column = candidate gene IDs) and `out_dir` (output directory); `../../../Gifu_all_samples.tpm.tsv` and `../../../MG20_all_samples.tpm.tsv`; `all_new_genes.list`, `SNF_gene.list` and the orthology table `All_species_merged_by_GifuT2T.tsv` in the working directory.
**Outputs**: `/SNF_vs_ALL/` and `/SNF_vs_New/`, each with `edges_bonferroni.txt`, `nodes.txt` and `network_igraph.rds`; `/network_gene_composition.txt` with the counts per node class.
**Run**: `Rscript co-exp_psych.R go_id_list.txt output_dir` (run from a directory three levels below the two TPM matrices).
**Tools**: R 4.4.3 with psych (`corr.test(method = "pearson", adjust = "none")`), igraph, data.table and tidyverse; settings as coded: sample columns matching `root_hpi48`/`root_uni`/`nodule`, minimum TPM > 0.5, `alpha = 0.05` with `p_cutoff = alpha / nrow(edges_all)`, edges kept when `p < p_cutoff` and `cor > 0.7`.