在R中为SpiecEasi网络顶点分配元数据的技术问询

步骤1：提取Substrata元数据 #

根据你的TreeSE对象结构，先把样本层面的substrata信息提取出来；如果是微生物共现网络，需要把substrata信息关联到每个微生物类群（比如按类群在不同substrata中的丰度主导性标记）。

# 从TreeSE提取样本元数据
sample_meta <- as.data.frame(colData(你的TreeSE对象名))
substrata_col <- sample_meta$substrata # 替换为你的substrata列名
names(substrata_col) <- rownames(sample_meta)

# 如果是微生物共现网络（顶点为属水平类群），计算每个类群的主导substrata
library(mia)
# 用你已聚合好的属水平TreeSE对象
mean_abund <- assay(你的属水平TreeSE对象, "relabundance") %>% 
  t() %>%
  as.data.frame() %>%
  cbind(substrata = substrata_col) %>%
  dplyr::group_by(substrata) %>%
  dplyr::summarise(dplyr::across(everything(), mean)) %>%
  tibble::column_to_rownames("substrata") %>%
  t() %>%
  as.data.frame()
# 为每个类群分配丰度最高的substrata标签
microbe_substrata <- apply(mean_abund, 1, function(x) names(x)[which.max(x)])
names(microbe_substrata) <- rownames(mean_abund)

步骤2：将SpiecEasi结果转为igraph并添加顶点属性 #

把SpiecEasi的拟合结果转成igraph对象，再把substrata信息绑定到顶点属性上：

library(igraph)
library(SpiecEasi)

# 从SpiecEasi结果中提取拟合好的邻接矩阵并转成igraph
se_igraph <- adj2igraph(getRefit(你的SpiecEasi结果对象))

# 绑定属性：分两种场景
# 场景1：网络顶点为样本（样本共现网络）
V(se_igraph)$substrata <- substrata_col[V(se_igraph)$name]

# 场景2：网络顶点为微生物类群（共现网络）
V(se_igraph)$substrata <- microbe_substrata[V(se_igraph)$name]

步骤3：高亮展示Substrata信息 #

用igraph绘图时，根据substrata属性给顶点着色，实现高亮：

# 定义颜色映射
substrata_types <- unique(V(se_igraph)$substrata)
color_pal <- RColorBrewer::brewer.pal(length(substrata_types), "Set2")
names(color_pal) <- substrata_types

# 绘制网络
plot(se_igraph,
     vertex.color = color_pal[V(se_igraph)$substrata],
     vertex.label = NA, # 顶点过多时可隐藏标签
     vertex.size = 7,
     edge.width = 0.4,
     main = "Microbial Network Colored by Substrata")

# 添加图例
legend("topright",
       legend = substrata_types,
       fill = color_pal,
       title = "Substrata",
       cex = 0.8)

额外：验证聚类与Substrata的关联及桥接物种分析 #

# 用Louvain算法识别网络聚类
se_communities <- cluster_louvain(se_igraph)

# 卡方检验聚类与substrata的相关性
cluster_sub_df <- data.frame(
  cluster = membership(se_communities),
  substrata = V(se_igraph)$substrata
)
chisq.test(table(cluster_sub_df$cluster, cluster_sub_df$substrata))

# 筛选桥接物种（基于betweenness中心性）
V(se_igraph)$betweenness <- betweenness(se_igraph)
top_bridge_species <- V(se_igraph)[order(V(se_igraph)$betweenness, decreasing = T)[1:10]]
# 查看桥接物种的substrata属性
data.frame(
  species = names(top_bridge_species),
  substrata = top_bridge_species$substrata,
  betweenness_score = top_bridge_species$betweenness
)

内容的提问来源于stack exchange，提问作者Jan