R中非结构化数据的Circos图答案

【问题标题】：Circos plots from unstructured data in RR中非结构化数据的Circos图
【发布时间】：2020-02-22 21:44:18
【问题描述】：

我有 100 多个观察点（点），在那里我查看了特定项目（例如 A、B、C..等）。数据的基本格式如下：

spot1<-c("A","B","B","B","C")
spot2<-c("C","NA","D","D","D")
spot3<-c("D","N","E","F","G")
spot4<-c("H","I","I","NA","NA")
spot5<-c("B","I","NA","NA","NA")
data1<-as.data.frame(cbind(spot1,spot2,spot3,spot4,spot5))
print(data1)

现在我想开发一个 adjacency matrix 的项目以使用 circlize 包在 R 中查看他们的网络。建议将不胜感激。

【问题讨论】：

标签： r neural-network adjacency-matrix circlize

【解决方案1】：

我会给你一个例子，如何使用 R 中的circlize 包绘制chord 图表：

生成随机数据：

set.seed(1000)
mat = matrix(sample(18, 18), 3, 6) 
rownames(mat) = paste0("S", 1:3)
colnames(mat) = paste0("E", 1:6)
print(mat)

输出

   E1 E2 E3 E4 E5 E6
S1  6 11  9 10 16 17
S2 13  8  7 12  5 14
S3  2  1  3 18  4 15

转换为adjacency矩阵：

df = data.frame(from = rep(rownames(mat), times = ncol(mat)),
    to = rep(colnames(mat), each = nrow(mat)),
    value = as.vector(mat),
    stringsAsFactors = FALSE)
print(df)

输出

   from to value
1    S1 E1     6
2    S2 E1    13
3    S3 E1     2
4    S1 E2    11
5    S2 E2     8
6    S3 E2     1
7    S1 E3     9
8    S2 E3     7
9    S3 E3     3
10   S1 E4    10
11   S2 E4    12
12   S3 E4    18
13   S1 E5    16
14   S2 E5     5
15   S3 E5     4
16   S1 E6    17
17   S2 E6    14
18   S3 E6    15

绘制chord图表：

chordDiagram(df, transparency = 0.5)

Adjacency matrix 你的数据：

代码：

spot1<-c("A","B","B","B","C")
spot2<-c("C","NA","D","D","D")
spot3<-c("D","N","E","F","G")
spot4<-c("H","I","I","NA","NA")
spot5<-c("B","I","NA","NA","NA")
data1<-as.data.frame(cbind(spot1,spot2,spot3,spot4,spot5))


mat<-data1
df = data.frame(from = rep(rownames(mat), times = ncol(mat)),
                to = rep(colnames(mat), each = nrow(mat)),
                value = as.vector(mat),
                stringsAsFactors = FALSE)
print(df)

输出

   from    to value.spot1 value.spot2 value.spot3 value.spot4 value.spot5
1     1 spot1           A           C           D           H           B
2     2 spot1           B          NA           N           I           I
3     3 spot1           B           D           E           I          NA
4     4 spot1           B           D           F          NA          NA
5     5 spot1           C           D           G          NA          NA
6     1 spot2           A           C           D           H           B
7     2 spot2           B          NA           N           I           I
8     3 spot2           B           D           E           I          NA
9     4 spot2           B           D           F          NA          NA
10    5 spot2           C           D           G          NA          NA
11    1 spot3           A           C           D           H           B
12    2 spot3           B          NA           N           I           I
13    3 spot3           B           D           E           I          NA
14    4 spot3           B           D           F          NA          NA
15    5 spot3           C           D           G          NA          NA
16    1 spot4           A           C           D           H           B
17    2 spot4           B          NA           N           I           I
18    3 spot4           B           D           E           I          NA
19    4 spot4           B           D           F          NA          NA
20    5 spot4           C           D           G          NA          NA
21    1 spot5           A           C           D           H           B
22    2 spot5           B          NA           N           I           I
23    3 spot5           B           D           E           I          NA
24    4 spot5           B           D           F          NA          NA
25    5 spot5           C           D           G          NA          NA

【讨论】：

谢谢@ashwin-agrawal 。我想知道从我的数据生成邻接矩阵的代码（即所有点的项目之间的连接数）/
我已经在你的数据上复制了上面的代码来生成adjacency matrix，请在上面查看。
其实我不需要点名，只需要项目（例如，A，BC....I）和所有点的项目之间的共现次数之和（例如，AB =3，BC=3，CD=3（单向））
从上面的邻接矩阵中你可以很容易地通过取from列的差来得到它。
如果解决了问题，请点赞并接受答案