效果如下：

    33. 实现漂亮点图的注释（圈起来） - 图1

    记录几个功能，首先是放大区域：

    1. ggplot(data_matrix, aes(gene, exon)) +
    2.   geom_point() + geom_smooth(method = "lm") + 
    3.   geom_point(data=subset( data_matrix,gene < 2.5 & exon > (0.936*gene + 0.4094)), 
    4.              color = "red") + 
    5.   facet_zoom(x = group == "marked")

    33. 实现漂亮点图的注释（圈起来） - 图2

    x, y 控制放大是在横坐标还是纵坐标。

    此外，还可以绘制选择图片的区域轮廓：

    参见，https://www.cnblogs.com/Mao1518202/p/11961495.html

    比如：

    1. ggplot(data_matrix, aes(gene, exon)) +
    2.   geom_point() + geom_smooth(method = "lm") + 
    3.   geom_point(data=subset( data_matrix,gene < 2.5 & exon > (0.936*gene + 0.4094)), 
    4.              color = "red") + 
    5.   geom_mark_ellipse(data=subset( data_matrix,gene < 2.5 & exon > (0.936*gene + 0.4094)), 
    6.                     aes(label = group), color = "red")

    33. 实现漂亮点图的注释（圈起来） - 图3

    但如果不是这样的连续性数据呢？

    比如分类的一个个散点（气泡图）：

    33. 实现漂亮点图的注释（圈起来） - 图4

    发现并不能很好的识别出相邻的数据（每个点都是单独的一个圈圈），只能实现上面的效果啦～

    1. my_test <- data.frame(Features = rep(paste0("CD", 1:20), 3),
    2.                       Identity = as.character(sample(10, 60 ,replace = T)),
    3.                       Averange_Expression = runif(60, -1, 5),
    4.                       Percent_Expression = runif(60, 0, 90))
    5. tmp <- data.frame(Features = paste0("CD", 1:20), 
    6.                   Big_Features = as.character(sample(LETTERS[1:6], 20 ,replace = T)))
    7. my_test <- merge(my_test, tmp, by = "Features")
    9. my_test$Labels <- case_when(
    10.   my_test$Features %in% c("CD13", "CD16") & my_test$Identity == "5" ~ "Big Cell",
    11.   my_test$Features %in% c("CD15", "CD5") & my_test$Identity == "6" ~ "Ultra Big Cell",
    12.   my_test$Features %in% c("CD1", "CD10", "CD18") & my_test$Identity == "1" ~ "Little Cell",
    13.   TRUE ~ "no-marked"
    14. )
    16. my_test[!my_test$Labels %in% "no-marked" & duplicated(my_test$Labels), ]$Labels <- NA
    18. ggplot(my_test) + geom_point(aes(Features, Identity, 
    19.                                  color = Averange_Expression,
    20.                                  size = Percent_Expression)) +
    21.   theme_bw() + facet_grid(.~Big_Features, scales = "free") + 
    22.   theme(axis.title.x = element_blank(),
    23.         axis.text.x = element_text(angle = 90),
    24.         text = element_text(size = 14),
    25.         strip.background = element_rect(fill = "transparent")) + 
    26.   geom_mark_rect(data = my_test[(!my_test$Labels %in% "no-marked") & (!is.na(my_test$Labels)), ],
    27.                  aes(Features, Identity, label = Labels), 
    28.                  color = "red") + 
    29.   geom_mark_rect(data = my_test[(!my_test$Labels %in% "no-marked") & (is.na(my_test$Labels)), ],
    30.                  aes(Features, Identity), 
    31.                  color = "red")

    有大神的话，可以帮帮忙～