查找范圍重疊的行

Question

我有 3 個大數據框，如下所示：

library(tibble)

df1 <- tibble(peak=c("peak1","peak2","peak3"), 
              coord1=c(100,500,1000),
              coord2=c(250,700,1250))


df2 <- tibble(peak=c("peak5","peak6","peak7"), 
              coord1=c(120,280,900),
              coord2=c(300,400,1850))


df3 <- tibble(peak=c("peak8","peak9","peak10"), 
              coord1=c(900,3000,5600),
              coord2=c(2000,3400,5850))

df1
#> # A tibble: 3 × 3
#>   peak  coord1 coord2
#>   <chr>  <dbl>  <dbl>
#> 1 peak1    100    250
#> 2 peak2    500    700
#> 3 peak3   1000   1250
df2
#> # A tibble: 3 × 3
#>   peak  coord1 coord2
#>   <chr>  <dbl>  <dbl>
#> 1 peak5    120    300
#> 2 peak6    280    400
#> 3 peak7    900   1850
df3
#> # A tibble: 3 × 3
#>   peak   coord1 coord2
#>   <chr>   <dbl>  <dbl>
#> 1 peak8     900   2000
#> 2 peak9    3000   3400
#> 3 peak10   5600   5850

我是 R 的新手，我試圖找到每個數據幀唯一的坐標 (coord1, coord2) 內的重疊區域、兩個數據幀之間的重疊以及所有數據幀內的重疊。

我想要這些數據框作為輸出。 目前我很難找到如何在 R、dplyr 中指定我想根據重疊范圍進行過濾。 我缺少一個命令

獨特的是這些峰的范圍不與其他數據幀的峰范圍重疊

> unique

peak    coord1  coord2
peak6    280     400
peak9    3000    3400
peak10   5600    5850

df1-df2 之間共有

>df1df2 
peak       coord1  coord2
peak1       100     250
peak5       120     300
peak3      1000    1250
peak7       900    1850

df1-df3 之間共有

peak       coord1  coord2
peak3       1000    1250
peak8       900   2000

然后在 df1-df2-df3 之間通用

Answer 1

老實說，我不明白你搜索的最終目標是什么。 無論如何，有一個解決方案使用ivs方法和 ivs 包中的函數來檢查向量的間隔。 這不是一個優雅的解決方案，它沒有考慮同一數據幀內的重疊向量。

# load packages
library(tidyverse)
library(ivs)

您的數據

df1 <- tibble(peak=c("peak1","peak2","peak3"), 
              coord1 = c(100, 500, 1000),
              coord2 = c(250, 700, 1250))


df2 <- tibble(peak=c("peak5","peak6","peak7"), 
              coord1 = c(120, 280, 900),
              coord2 = c(300, 400, 1850))


df3 <- tibble(peak=c("peak8","peak9","peak10"), 
              coord1 = c(900, 3000, 5600),
              coord2 = c(2000, 3400, 5850))

使用函數iv_overlaps來創建間隔

check_df1_df2 <- df1 %>%
      mutate(any_overlap = iv_overlaps(range, df2$range),
             check = "df1-df2")

check_df1_df3 <- df1 %>%
      mutate(any_overlap = iv_overlaps(range, df3$range),
             check = "df1-df3")

check_df2_df1 <- df2 %>%
      mutate(any_overlap = iv_overlaps(range, df1$range),
             check = "df2-df1")

check_df2_df3 <- df2 %>%
      mutate(any_overlap = iv_overlaps(range, df3$range),
             check = "df2-df3")

check_df3_df1 <- df3 %>%
      mutate(any_overlap = iv_overlaps(range, df1$range),
             check = "df3-df1")

check_df3_df2 <- df3 %>%
      mutate(any_overlap = iv_overlaps(range, df2$range),
             check = "df3-df2")

綁定數據框

final_conclusion <- bind_rows(check_df1_df2, check_df1_df3, check_df2_df1, check_df2_df3, check_df3_df1, check_df3_df2, .id = "df_check") %>% 
      group_by(peak) %>% 
      mutate(overlapping_sum = sum(any_overlap))

檢查數據幀之間的重疊間隔

overlapping <- final_conclusion %>% 
      filter(overlapping_sum > 0) %>% 
      pivot_wider(id_cols = peak, names_from = check, values_from = range)

> overlapping

# A tibble: 5 × 7
# Groups:   peak [5]
  peak     `df1-df2`    `df1-df3`   `df2-df1`   `df2-df3`   `df3-df1`   `df3-df2`
  <chr>    <iv<dbl>>    <iv<dbl>>   <iv<dbl>>   <iv<dbl>>   <iv<dbl>>   <iv<dbl>>
1 peak1   [100, 250)   [100, 250)    [NA, NA)    [NA, NA)    [NA, NA)    [NA, NA)
2 peak3 [1000, 1250) [1000, 1250)    [NA, NA)    [NA, NA)    [NA, NA)    [NA, NA)
3 peak5     [NA, NA)     [NA, NA)  [120, 300)  [120, 300)    [NA, NA)    [NA, NA)
4 peak7     [NA, NA)     [NA, NA) [900, 1850) [900, 1850)    [NA, NA)    [NA, NA)
5 peak8     [NA, NA)     [NA, NA)    [NA, NA)    [NA, NA) [900, 2000) [900, 2000)

檢查數據幀之間的非重疊間隔

not_overlapping <- final_conclusion %>% 
      filter(overlapping_sum == 0) %>% 
      pivot_wider(id_cols = peak, names_from = check, values_from = range)

> not_overlapping
# A tibble: 4 × 7
# Groups:   peak [4]
  peak    `df1-df2`  `df1-df3`  `df2-df1`  `df2-df3`    `df3-df1`    `df3-df2`
  <chr>   <iv<dbl>>  <iv<dbl>>  <iv<dbl>>  <iv<dbl>>    <iv<dbl>>    <iv<dbl>>
1 peak2  [500, 700) [500, 700)   [NA, NA)   [NA, NA)     [NA, NA)     [NA, NA)
2 peak6    [NA, NA)   [NA, NA) [280, 400) [280, 400)     [NA, NA)     [NA, NA)
3 peak9    [NA, NA)   [NA, NA)   [NA, NA)   [NA, NA) [3000, 3400) [3000, 3400)
4 peak10   [NA, NA)   [NA, NA)   [NA, NA)   [NA, NA) [5600, 5850) [5600, 5850)