5步将凌乱的数据处理成干净的表格。

library(tidyverse)
library(here)
library(scales)

1 第一步：记住几个重要参数的清单

字体：Arial 或 Helvetica，正文11pt，标题12pt。
主题：黑白主题，theme_classic()，无网格背景。
分辨率：300 DPI以上。
格式：TIFF（首选）或 EPS。
尺寸：单栏图约3.5英寸，双栏图约7.5英寸。
配色：适合黑白印刷（灰度配色）或色盲友好配色。
图例：优先放图内，不在图外单独占空间。

把这个清单保存下来，每次画图的时候粘进提示词，DeepSeek V4给你的代码就直接是投稿级别的。

2 第二步：通用的图标导出函数

每次画完图都要导出，把这个函数放在脚本最前面，之后直接调用：

# 投稿级的图表导出函数
save_figure <- function(
  plot,
  filename,
  width = 7,
  height = 5,
  dpi = 300,
  device = "tiff"
) {
  ggsave(
    filename = here("outputs", "figures", paste0(filename, ".", device)),
    plot = plot,
    width = width,
    height = height,
    units = "in",
    dpi = dpi,
    device = device
  )
  message("✅ 图表已保存：", filename, ".", device)
}

3 第三步：投稿的通用主题theme

把这个主题对象放在脚本里，每张图都加上+ theme_publication()：

theme_publication <- function(base_size = 11, base_family = "Arial") {
  theme_classic(base_size = base_size, base_family = base_family) +
    theme_sub_legend(
      title = element_text(size = base_size, face = "bold"),
      text = element_text(size = base_size - 1),
      background = element_rect(
        fill = "white",
        color = "grey80",
        linewidth = 0.3
      ),
      key.size = unit(0.4, "cm")
    ) +
    theme_sub_plot(
      title = element_text(
        size = base_size + 1,
        face = "bold",
        hjust = 0.5
      ),
      margin = margin(10, 15, 10, 10),
      background = element_rect(fill = "white", color = NA),
    ) +
    theme_sub_axis(
      title = element_text(size = base_size, face = "bold"),
      text = element_text(size = base_size - 1, color = "black")
    ) +
    theme_sub_panel(
      background = element_rect(fill = "white", color = NA)
    ) +
    theme(
      # 文字
      text = element_text(family = base_family, size = base_size),
    )
}

4 散点图

最常用场景：探索两个连续变量之间的关系，展示相关性，添加回归线。

mtcars_df <- mtcars |>
  rownames_to_column("car") |>
  mutate(cyl = factor(cyl, labels = c("4 缸", "6 缸", "8 缸")))
head(mtcars_df)

                car  mpg  cyl disp  hp drat    wt  qsec vs am gear carb
1         Mazda RX4 21.0 6 缸  160 110 3.90 2.620 16.46  0  1    4    4
2     Mazda RX4 Wag 21.0 6 缸  160 110 3.90 2.875 17.02  0  1    4    4
3        Datsun 710 22.8 4 缸  108  93 3.85 2.320 18.61  1  1    4    1
4    Hornet 4 Drive 21.4 6 缸  258 110 3.08 3.215 19.44  1  0    3    1
5 Hornet Sportabout 18.7 8 缸  360 175 3.15 3.440 17.02  0  0    3    2
6           Valiant 18.1 6 缸  225 105 2.76 3.460 20.22  1  0    3    1

# 基础散点图
p_scatter_basic <- ggplot(
  mtcars_df,
  aes(x = wt, y = mpg)
) +
  geom_point(aes(color = cyl, shape = cyl), size = 2.5, alpha = 0.8) +
  geom_smooth(
    aes(group = 1),
    method = "lm",
    se = T,
    color = "black",
    linewidth = 0.8,
    alpha = 0.15
  )
p_scatter_basic

# 加入投稿级配色，应用投稿主题
p_scatter <- p_scatter_basic +
  ggsci::scale_color_lancet() + # ggsici提供的Lancet配色方案，适合科学图表
  theme_publication() +
  theme(
    legend.position = c(0.85, 0.85) # 将图例放在图内右上角
  ) +
  labs(
    x = "车重 (wt)",
    y = "每加仑英里数 (mpg)",
    color = "气缸数",
    shape = "气缸数",
    title = NULL
  )
p_scatter

# 导出图表
save_figure(p_scatter, "fig1_scatter")

5 箱线图

最常用场景：比较多个组的连续变量分布，展示中位数、四分位数和异常值。

p_boxplot_basic <- ggplot(
  iris,
  aes(x = Species, y = Sepal.Length, fill = Species)
) +
  geom_boxplot(
    alpha = 0.7,
    outlier.shape = 21, # 使用空心圆表示异常值
    outlier.size = 1.5,
    outlier.alpha = 0.6,
    width = 0.5
  ) +
  # 添加散点以显示数据分布，使用jitter避免重叠
  geom_jitter(
    width = 0.15,
    alpha = 0.4,
    size = 1
  )
p_boxplot_basic

# 加入投稿级配色，应用投稿主题
p_boxplot <- p_boxplot_basic +
  ggsci::scale_fill_lancet() + # 使用Lancet配色方案
  theme_publication() +
  theme(
    legend.position = "none" # 分组已经在x轴和颜色显示，不显示图例
  ) +
  labs(
    x = "物种",
    y = "花萼长度 (cm)",
    title = NULL
  )
p_boxplot

# 在箱线图中加入统计显著性标记
p_boxplot_stat <- p_boxplot +
  ggpubr::stat_compare_means(
    method = "wilcox.test",
    comparisons = list(
      c("setosa", "versicolor"),
      c("setosa", "virginica"),
      c("versicolor", "virginica")
    ),
    label = "p.signif", # p.singificant显示星号，p.format显示具体p值
    label.y = c(7, 7.5, 8), # 调整显著性标记的位置
    step.increase = 0.1 # 每个比较之间增加10%的高度，避免重叠
  )
p_boxplot_stat

# 导出图表
save_figure(p_boxplot_stat, "fig2_boxplot")

6 柱状图

最常用场景：比较不同组的均值或总和，展示分类变量的分布。

# lable: bar-plot
p_bar_df <- iris |> 
  group_by(Species) |> 
  summarise(
    mean_sl = mean(Sepal.Length),
    se_sl = sd(Sepal.Length) / sqrt(n())
  )

# 基础柱状图--显示均值和误差线
p_bar_basic <- ggplot(p_bar_df, aes(x = Species, y = mean_sl, fill = Species)) +
  geom_col(width = 0.6, alpha = 0.5) +
  # 误差棒
  geom_errorbar(
    aes(ymin = mean_sl - se_sl, ymax = mean_sl + se_sl),
    width = 0.2,
    # color = "black",
    linewidth = 0.7
  )
p_bar_basic

# 加入投稿级配色，应用投稿主题
p_bar <- p_bar_basic +
  ggsci::scale_fill_lancet() + # 使用Lancet配色方案
  theme_publication() +
  theme(
    legend.position = "none" # 分组已经在x轴和颜色显示，不显示图例
  ) +
  scale_y_continuous(expand = expansion(mult = c(0, 0.1))) + # 调整y轴起点为0，增加10%空间
  labs(
    x = "物种",
    y = "花萼长度均值 (cm ± SE)",
    title = NULL
  )
p_bar

# 导出图表
save_figure(p_bar, "fig3_bar")

7 折线图

最常用场景：展示随时间或其他连续变量变化的趋势，多组对比。

# 构造示例数据
set.seed(42)
followup_data <- expand.grid(
  time = c(0, 3, 6, 9, 12),
  group = c("治疗组", "对照组")
) |>
  mutate(
    mean_score = c(
      50,
      45,
      38,
      30,
      25, # 治疗组: time 0,3,6,9,12
      50,
      49,
      48,
      47,
      46
    ), # 对照组: time 0,3,6,9,12
    se = c(3, 3, 2.5, 2, 2, 3, 3, 3, 3, 3)
  )

# 基础折线图-显示置信区间
p_line_basic <- ggplot(
  followup_data,
  aes(x = time, y = mean_score, color = group, shape = group)
) +
  geom_line(linewidth = 0.9) +
  geom_point(size = 3) +
  geom_ribbon( # 添加置信区间
    aes(ymin = mean_score - se, ymax = mean_score + se, fill = group),
    alpha = 0.15,
    color = NA
  )
p_line_basic

# 加入投稿级配色，应用投稿主题
p_line <- p_line_basic +
  ggsci::scale_color_lancet() + # 使用Lancet配色方案
  ggsci::scale_fill_lancet() + # 填充颜色与线条颜色一致
  theme_publication() +
  theme(
    legend.position = c(0.85, 0.5) # 将图例放在图内右上角
  ) +
  labs(
    x = "随访时间 (月)",
    y = "平均评分 (分值 ± SE)",
    color = "治疗方案",
    shape = "治疗方案",
    fill = "治疗方案",
    title = NULL
  )
p_line

# 导出图表
save_figure(p_line, "fig4_line")

8 热图

最常用场景：相关矩阵可视化，基因表达热图，样本聚类分析。

# 使用burceR
p_heatmap_basic <- bruceR::Corr(
  mtcars |> select(mpg, cyl, disp, hp, wt, qsec),
  method = "pearson",
)$plot

Pearson's r and 95% confidence intervals:
────────────────────────────────────────────
               r       [95% CI]     p      N
────────────────────────────────────────────
mpg-cyl    -0.85 [-0.93, -0.72] <.001 *** 32
mpg-disp   -0.85 [-0.92, -0.71] <.001 *** 32
mpg-hp     -0.78 [-0.89, -0.59] <.001 *** 32
mpg-wt     -0.87 [-0.93, -0.74] <.001 *** 32
mpg-qsec    0.42 [ 0.08,  0.67]  .017 *   32
cyl-disp    0.90 [ 0.81,  0.95] <.001 *** 32
cyl-hp      0.83 [ 0.68,  0.92] <.001 *** 32
cyl-wt      0.78 [ 0.60,  0.89] <.001 *** 32
cyl-qsec   -0.59 [-0.78, -0.31] <.001 *** 32
disp-hp     0.79 [ 0.61,  0.89] <.001 *** 32
disp-wt     0.89 [ 0.78,  0.94] <.001 *** 32
disp-qsec  -0.43 [-0.68, -0.10]  .013 *   32
hp-wt       0.66 [ 0.40,  0.82] <.001 *** 32
hp-qsec    -0.71 [-0.85, -0.48] <.001 *** 32
wt-qsec    -0.17 [-0.49,  0.19]  .339     32
────────────────────────────────────────────

# 加入投稿级配色，应用投稿主题
p_heatmap <- p_heatmap_basic +
  theme_publication()
p_heatmap

# 导出图表
save_figure(p_heatmap, "fig5_heatmap")