如何用pairwise.adonis2针对特定因子水平做成对PERMANOVA比较？

要分别在10m深度和50m深度下检验2020年与2023年的群落差异，可通过拆分数据集或批量分组分析实现，以下是具体方案：

准备工作（优化数据处理） #

先将原始数据中的分类变量转换为因子类型，避免数值型干扰：

library(tidyverse)
library(vegan)
library(pairwiseAdonis)

# 原始数据
df <- structure(list(Sample = c("A1", "A2", "A3", "A4", "A5", "B1", 
"B2", "B3", "B4", "B5", "C1", "C2", "C3", "C4", "C5", "D1", "D2", 
"D3", "D4", "D5"), Depth = c(10, 10, 10, 10, 10, 50, 50, 50, 
50, 50, 10, 10, 10, 10, 10, 50, 50, 50, 50, 50), Year = c(2020, 
2020, 2020, 2020, 2020, 2020, 2020, 2020, 2020, 2020, 2023, 2023, 
2023, 2023, 2023, 2023, 2023, 2023, 2023, 2023), Sp1 = c(0, 0, 
0, 0, 1, 0, 5, 0, 0, 8, 0, 6, 0, 4, 0, 12, 16, 6, 8, 4), Sp2 = c(3, 
7, 9, 2, 6, 4, 5, 9, 4, 3, 6, 8, 6, 4, 7, 9, 2, 5, 3, 3), Sp3 = c(1, 
0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0), Sp4 = c(11, 
32, 12, 17, 5, 0, 13, 19, 3, 4, 5, 0, 24, 20, 15, 3, 2, 0, 1, 
8), Sp5 = c(5, 6, 7, 2, 3, 4, 8, 1, 1, 1, 0, 12, 6, 3, 8, 6, 
2, 0, 1, 2), Sp6 = c(0, 0, 0, 0, 0, 5, 6, 7, 3, 6, 0, 0, 0, 0, 
0, 15, 12, 4, 8, 11)), row.names = c(NA, -20L), class = "data.frame")

# 转换分类变量为因子类型
df_combined <- df %>%
  mutate(Depth = as.factor(Depth), Year = as.factor(Year))

方法一：手动拆分数据集单独分析 #

分别提取10m和50m深度的子集，独立运行PERMANOVA：

# 处理10m深度数据
depth_10 <- df_combined %>% filter(Depth == 10)
species_10 <- depth_10 %>% select(starts_with("Sp"))
factors_10 <- depth_10 %>% select(Sample, Depth, Year)

# 10m深度下年份对群落的PERMANOVA
adonis_depth10 <- adonis2(species_10 ~ Year, data = factors_10, permutations = 9999, method = "bray")
cat("10m深度下年份的PERMANOVA结果：\n")
print(adonis_depth10)

# 处理50m深度数据
depth_50 <- df_combined %>% filter(Depth == 50)
species_50 <- depth_50 %>% select(starts_with("Sp"))
factors_50 <- depth_50 %>% select(Sample, Depth, Year)

# 50m深度下年份对群落的PERMANOVA
adonis_depth50 <- adonis2(species_50 ~ Year, data = factors_50, permutations = 9999, method = "bray")
cat("\n50m深度下年份的PERMANOVA结果：\n")
print(adonis_depth50)

方法二：分组批量处理（高效） #

用dplyr的分组功能批量运行各深度下的PERMANOVA，适合多水平场景：

# 按Depth分组，批量执行PERMANOVA
depth_year_adonis <- df_combined %>%
  group_by(Depth) %>%
  group_modify(function(data, key) {
    species_data <- data %>% select(starts_with("Sp"))
    adonis_result <- adonis2(species_data ~ Year, data = data, permutations = 9999, method = "bray")
    # 将结果转换为数据框，统一输出
    as.data.frame(adonis_result) %>% mutate(Depth = key$Depth)
  }) %>%
  ungroup()

cat("各深度下年份的PERMANOVA批量结果：\n")
print(depth_year_adonis)

成对比较（含多重检验控制） #

若需更细致的年份成对差异检验，可在每个深度子集内运行pairwise.adonis2：

# 10m深度下年份成对比较
pairwise_depth10 <- pairwise.adonis2(species_10 ~ Year, data = factors_10, permutations = 9999)
cat("\n10m深度下年份的成对比较结果：\n")
print(pairwise_depth10)

# 50m深度下年份成对比较
pairwise_depth50 <- pairwise.adonis2(species_50 ~ Year, data = factors_50, permutations = 9999)
cat("\n50m深度下年份的成对比较结果：\n")
print(pairwise_depth50)

通过对比两个深度的PERMANOVA结果（R²值、P值），即可判断不同深度的群落受年份影响的差异。

内容的提问来源于stack exchange，提问作者mbroadribb