写在前面
准备数据集
直接开始绘图
library(Hmisc)library(dplyr)library(linkET)library(ggplot2)df <- read.csv(file.choose(), header = TRUE, row.names = 1)target_factors <- c("AGB", "BGB")env_factors <- setdiff(colnames(df), target_factors)
# 分离环境因子数据env <- df[, env_factors]# 计算相关性矩阵和 P 值矩阵df_rcorr <- rcorr(as.matrix(df), type = "pearson")r <- df_rcorr$r # 相关系数矩阵p <- df_rcorr$P # P 值矩阵# 初始化结果数据框results <- data.frame(P = character(), R = numeric(), Y = character(), X = character(), sig = character(), type = character(), stringsAsFactors = FALSE)# 遍历每个目标因子和环境因子的组合，计算相关性for (y in target_factors) {for (x in env_factors) {r_value <- r[x, y]p_value <- p[x, y]# 确定显著性水平和相关性方向significance <- ifelse(p_value < 0.05, "< 0.05", ">= 0.05")sig <- ifelse(r_value > 0, ">0", "<0")type <- ifelse(r_value > 0, "实线", "虚线")# 添加结果到数据框results <- rbind(results, data.frame(P = significance, R = round(r_value, 2), Y = y, X = x, sig = sig, type = type))}}# 查看结果print(results)# 设置颜色映射cols <- c(">= 0.05" = "grey", "< 0.05" = "#1B9E77", "< 0.01" = "#D95F02")# 绘制相关性图，带显著性标记qcorrplot(correlate(env), type = "lower", diag = FALSE) +geom_square() +geom_text(aes(label = ifelse(p < 0.001, "***",ifelse(p < 0.01, "**",ifelse(p < 0.05, "*", "")))),size = 5, color = "black") + # 根据 p 值设置星号geom_couple(aes(colour = P, # 使用显著性 P 值来设置线的颜色size = abs(R), # 根据相关系数 R 的绝对值调整线条粗细linetype = type, # 根据 type 列选择实线或虚线from = Y, to = X # from 和 to 指定因变量和环境因子), data = results, curvature = 0.15) +scale_fill_gradientn(colours = RColorBrewer::brewer.pal(3, "RdBu")) +scale_size_continuous(range = c(0.1, 1)) + # 设置线条粗细范围scale_colour_manual(values = cols) +scale_linetype_manual(values = c("实线" = "solid", "虚线" = "dashed")) + # 设置线型映射guides(size = guide_legend(title = "Correlation Strength",override.aes = list(colour = "grey35"), order = 2),colour = guide_legend(title = "P value",override.aes = list(size = 3), order = 1),fill = guide_colorbar(title = "Pearson's r", order = 3),linetype = guide_legend(title = "Line Type", order = 4)) +theme_minimal() +theme(panel.grid = element_blank(), # 移除网格axis.text.x = element_text(angle = 45, hjust = 1)) + # x轴标签旋转labs(title = "Environmental Factors Correlation with Biomass",x = "Environmental Factor (X)", y = "Target Factor (Y)")
完整版代码如下
# 加载所需库library(Hmisc)library(dplyr)library(linkET)library(ggplot2)# 读取数据df <- read.csv(file.choose(), header = TRUE, row.names = 1)# 提取目标因子和环境因子target_factors <- c("AGB", "BGB")env_factors <- setdiff(colnames(df), target_factors)# 分离环境因子数据env <- df[, env_factors]# 计算相关性矩阵和 P 值矩阵df_rcorr <- rcorr(as.matrix(df), type = "pearson")r <- df_rcorr$r # 相关系数矩阵p <- df_rcorr$P # P 值矩阵# 初始化结果数据框results <- data.frame(P = character(), R = numeric(), Y = character(), X = character(), sig = character(), type = character(), stringsAsFactors = FALSE)# 遍历每个目标因子和环境因子的组合，计算相关性for (y in target_factors) {for (x in env_factors) {r_value <- r[x, y]p_value <- p[x, y]# 确定显著性水平和相关性方向significance <- ifelse(p_value < 0.05, "< 0.05", ">= 0.05")sig <- ifelse(r_value > 0, ">0", "<0")type <- ifelse(r_value > 0, "实线", "虚线")# 添加结果到数据框results <- rbind(results, data.frame(P = significance, R = round(r_value, 2), Y = y, X = x, sig = sig, type = type))}}# 查看结果print(results)# 设置颜色映射cols <- c(">= 0.05" = "grey", "< 0.05" = "#1B9E77", "< 0.01" = "#D95F02")# 绘制相关性图，带显著性标记qcorrplot(correlate(env), type = "lower", diag = FALSE) +geom_square() +geom_text(aes(label = ifelse(p < 0.001, "***",ifelse(p < 0.01, "**",ifelse(p < 0.05, "*", "")))),size = 5, color = "black") + # 根据 p 值设置星号geom_couple(aes(colour = P, # 使用显著性 P 值来设置线的颜色size = abs(R), # 根据相关系数 R 的绝对值调整线条粗细linetype = type, # 根据 type 列选择实线或虚线from = Y, to = X # from 和 to 指定因变量和环境因子), data = results, curvature = 0.15) +scale_fill_gradientn(colours = RColorBrewer::brewer.pal(3, "RdBu")) +scale_size_continuous(range = c(0.1, 1)) + # 设置线条粗细范围scale_colour_manual(values = cols) +scale_linetype_manual(values = c("实线" = "solid", "虚线" = "dashed")) + # 设置线型映射guides(size = guide_legend(title = "Correlation Strength",override.aes = list(colour = "grey35"), order = 2),colour = guide_legend(title = "P value",override.aes = list(size = 3), order = 1),fill = guide_colorbar(title = "Pearson's r", order = 3),linetype = guide_legend(title = "Line Type", order = 4)) +theme_minimal() +theme(panel.grid = element_blank(), # 移除网格axis.text.x = element_text(angle = 45, hjust = 1)) + # x轴标签旋转labs(title = "Environmental Factors Correlation with Biomass",x = "Environmental Factor (X)", y = "Target Factor (Y)")
★
转载自生态R学社
文章仅代表作者观点，与本公众号无关，版权归原作者所有
原文标题：相关性组合图最终版
·END·
图文编辑：王淑巾 武园伊
审编：周笑语 闫孟蕊
终审：初明若 李雨竹 代浩宇 毕丝淇
猜你喜欢
1.佳文赏析 | 李立华等：创造力空间：重新认识文化和旅游的本质
4.地学招聘 | 新奥尔良大学规划与城市研究系田光副教授招收全奖博士生
