ggplot绘制envfit结果时箭头与样点比例失调问题求助

我参考了很多用ggplot绘制metaMDS和envfit结果的教程及已解决案例，但按步骤操作后，仍出现变量箭头异常过长、完全掩盖样点图的问题。多次尝试用ordiArrowMul()缩放箭头，结果依旧无效。原生plot()函数绘图结果正常，仅ggplot版本出问题。

问题原因 #

ordiArrowMul(ef)返回的缩放系数是为原生plot()设计的，ggplot的坐标系统与原生plot存在差异，直接相乘会导致箭头过度放大。同时，部分环境变量的数值范围差异较大，也会加剧箭头长度失衡的问题。

修改后的代码 #

1. 数据读取与基础分析（保留原逻辑） #

# 读取数据
data <- read.csv('~/data.csv', row.names = 1) 
envi <- read.csv('~/env.csv', row.names = 1)

meta <- structure(list(SAMPLE_ID = c("MOLA_1-1_A", "MOLA_1-1_A", "MOLA_1-1_A", "MOLA_10-1_A", "MOLA_10-1_A", "MOLA_10-1_A", "MOLA_12-1_A", "MOLA_12-1_A", "MOLA_12-1_A", "MOLA_13-1_A", "MOLA_13-1_A", "MOLA_15-1_A", "MOLA_15-1_A", "MOLA_15-1_A", "MOLA_18-1_A", "MOLA_18-1_A", "MOLA_18-1_A", "MOLA_19-1_A", "MOLA_19-1_A", "MOLA_19-1_A", "MOLA_21-1_B", "MOLA_21-1_B", "MOLA_22-1_A", "MOLA_22-1_A", "MOLA_22-1_A", "MOLA_23-1_B", "MOLA_23-1_B", "MOLA_23-1_B", "MOLA_25-1_A", "MOLA_25-1_A", "MOLA_25-1_A", "MOLA_4-1_A", "MOLA_4-1_A", "MOLA_4-1_A", "MOLA_7-1_A", "MOLA_7-1_A", "MOLA_7-1_A", "MOLA_9-1_A", "MOLA_9-1_A", "MOLA_9-1_A"), `Month-Year` = structure(c(1556668800, 1556668800, 1556668800, 1590969600, 1590969600, 1590969600, 1593561600, 1593561600, 1593561600, 1596240000, 1596240000, 1604188800, 1604188800, 1604188800, 1612137600, 1612137600, 1612137600, 1614556800, 1614556800, 1614556800, 1617235200, 1617235200, 1619827200, 1619827200, 1619827200, 1625097600, 1625097600, 1625097600, 1564617600, 1564617600, 1564617600, 1559347200, 1559347200, 1559347200, 1569888000, 1569888000, 1569888000, 1577836800, 1577836800, 1577836800), tzone = "UTC", class = c("POSIXct", "POSIXt")), season = c("Spring", "Spring", "Spring", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Autumn", "Autumn", "Autumn", "Winter", "Winter", "Winter", "Spring", "Spring", "Spring", "Spring", "Spring", "Spring", "Spring", "Spring", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Summer", "Autumn", "Autumn", "Autumn", "Winter", "Winter", "Winter"), Depth = c("5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m", "5 m"), month = c("May", "May", "May", "Jun", "Jun", "Jun", "Jul", "Jul", "Jul", "Aug", "Aug", "Nov", "Nov", "Nov", "Feb", "Feb", "Feb", "Mar", "Mar", "Mar", "Apr", "Apr", "May", "May", "May", "Jul", "Jul", "Jul", "Aug", "Aug", "Aug", "Jun", "Jun", "Jun", "Oct", "Oct", "Oct", "Jan", "Jan", "Jan")), class = "data.frame", row.names = c(NA, -40L))

# NMDS分析
set.seed(123) 
nmds <- metaMDS(as.matrix(data), distance = "bray")

# envfit环境因子拟合
set.seed(123) 
ef <- envfit(nmds, envi, permu = 999) 

2. 优化后的ggplot绘图代码 #

# 提取显著环境因子箭头数据（弃用ordiArrowMul，改用自定义缩放）
en_coord_cont <- as.data.frame(scores(ef, "vectors"))
en_coord_cont$pval <- ef[["vectors"]][["pvals"]]
en_coord_cont <- dplyr::filter(en_coord_cont, pval <= 0.05)

# 自定义缩放系数：基于NMDS样点坐标范围调整箭头长度
max_nmds1 <- max(abs(scores(nmds)$NMDS1))
max_nmds2 <- max(abs(scores(nmds)$NMDS2))
scale_factor <- min(max_nmds1, max_nmds2) / 4  # 分母可按需调整（3-5之间）

# 对箭头坐标进行缩放
en_coord_cont <- en_coord_cont * scale_factor

# 提取样点NMDS得分并关联元数据
data.scores <- as.data.frame(scores(nmds)) 
temp <- cbind(data.scores, meta) 
temp$month <- factor(temp$month, levels = month.abb)

# ggplot绘图
xx <- ggplot(temp, aes(x = NMDS1, y = NMDS2)) +
  geom_point(size = 4, aes(shape = Depth, colour = month)) +
  coord_fixed() +
  theme(
    axis.text.y = element_text(colour = "black", size = 12, face = "bold"),
    axis.text.x = element_text(colour = "black", face = "bold", size = 12),
    legend.text = element_text(size = 12, face = "bold", colour = "black"),
    legend.position = "right",
    axis.title.y = element_text(face = "bold", size = 14),
    axis.title.x = element_text(face = "bold", size = 14, colour = "black"),
    legend.title = element_text(size = 14, colour = "black", face = "bold"),
    panel.background = element_blank(),
    panel.border = element_rect(colour = "black", fill = NA, size = 1.2),
    legend.key = element_blank()
  ) +
  scale_colour_manual(values = c('blue4','blue','green4','green3','green','red4','orangered','orange','yellow3','yellow')) +
  geom_vline(xintercept = 0, linetype = "dashed") +
  geom_hline(yintercept = 0, linetype = "dashed") +
  # 绘制缩放后的箭头
  geom_segment(
    aes(x = 0, y = 0, xend = NMDS1, yend = NMDS2),
    data = en_coord_cont,
    size = 1, alpha = 0.5, colour = "grey30"
  ) +
  # 文本标签偏移到箭头外侧，避免重叠
  geom_text(
    data = en_coord_cont,
    aes(x = NMDS1 * 1.1, y = NMDS2 * 1.1),
    colour = "grey30", fontface = "bold",
    label = row.names(en_coord_cont)
  )

print(xx)

关键调整说明 #

• 核心修改：放弃ordiArrowMul(ef)，改用基于NMDS样点坐标范围的自定义缩放系数，确保箭头长度与样点图比例匹配。
• 微调方式：若箭头长度仍不符合预期，可调整scale_factor的分母（比如改成3或5），按需适配。
• 细节优化：将环境变量标签偏移到箭头外侧，避免与箭头或样点重叠。

内容的提问来源于stack exchange，提问作者user18309655