rdataframematrix

Convert dataframe to matrix and create chord diagram

I am trying to create a chord diagram graph such as the one below.

enter image description here

Here, you can see which values of TRAV (such as TRAV21, TRAV1-2 etc) are matched with values of TRBV (such as TRBV6-4, TRBV6-1 ec), along with how often that pairing occurs.

Using the documentation provided (https://jokergoo.github.io/circlize_book/book/the-chorddiagram-function.html), I have to first create a matrix of my data in this format:

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

Then, convert the matrix to a dataframe in this format:

   from to value
 1    S1 E1     4
 2    S2 E1     7
 3    S3 E1     9
 4    S1 E2    14
 5    S2 E2     1
 6    S3 E2    10
 7    S1 E3    13
 8    S2 E3     6
 9    S3 E3     3
 10   S1 E4    17

So far, I have the code below:

df<- structure(list(TRAV = c("TRAV1-2", "TRAV1-2", "TRBV20-1", 
"TRAV12-1", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV8-2", "TRAV1-2", 
"TRBV20-1", "TRAV1-2", "TRAV1-2", "TRAV12-2", "TRAV12-3", "TRAV12-2", 
"TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV19", "TRAV1-2", "TRAV1-2", 
"TRAV19", "TRAV16", "TRAV27", "TRAV5", "TRAV1-2", "TRAV22", "TRAV1-2", 
"TRAV27", "TRAV26-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", 
"TRAV1-2", "TRAV41", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", 
"TRAV1-2", "TRAV35", "TRAV1-2", "TRAV1-2", "TRBV19", "TRAV1-2", 
"TRAV12-2", "TRAV1-2", "TRAV1-2", "TRAV16", "TRAV17", "TRAV35", 
"TRAV1-2", "TRBV4-1", "TRAV1-2", "TRBV5-5", "TRAV1-2", "TRBV6-4", 
"TRAV12-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV22", 
"TRAV1-2", "TRAV1-2", "TRAV8-4", "TRAV1-2", "TRAV1-2", "TRAV8-3", 
"TRBV5-1", "TRAV12-2", "TRAV1-2", "TRBV6-2", "TRAV19", "TRAV1-2", 
"TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV1-2", "TRAV21", 
"TRBV15", "TRAV24", "TRBV6-1", "TRAV1-2", "TRAV1-2", "TRAV1-2", 
"TRAV1-2", "TRAV12-2", "TRAV12-2", "TRAV29/DV5", "TRAV8-2", "TRAV12-2", 
"TRAV1-2", "TRAV1-2", "TRAV12-2", "TRAV1-2", "TRBV6-1", "TRAV1-2"
), TRBV = c("TRBV6-4", "TRBV6-2", NA, "TRBV4-3", "TRBV6-4", 
"TRBV6-2", "TRAV5", "TRBV6-2", "TRBV6-4", NA, "TRBV3-1", "TRBV6-2", 
"TRBV6-6", "TRBV10-2", "TRBV6-2", "TRBV6-4", "TRBV6-4", "TRBV6-4", 
"TRBV6-6", "TRBV20-1", "TRBV4-2", "TRBV9", "TRBV6-2", "TRBV5-5", 
"TRBV4-3", NA, "TRBV11-3", "TRBV4-2", "TRBV5-5", "TRBV9", "TRBV6-4", 
"TRBV6-4", "TRBV4-2", "TRBV4-3", "TRBV6-1", "TRBV12-4", "TRBV6-4", 
"TRBV6-4", "TRBV6-4", "TRBV6-4", "TRBV6-4", "TRBV12-4", "TRBV6-4", 
"TRBV6-4", NA, "TRBV6-4", "TRBV6-2", "TRBV19", "TRBV28", "TRBV20-1", 
"TRBV5-5", "TRAV41", "TRBV6-4", NA, "TRBV6-4", NA, "TRBV6-4", 
NA, "TRAV5", "TRAV23/DV6", "TRBV28", "TRBV6-4", "TRBV4-2", "TRAV35", 
"TRBV6-4", "TRBV6-4", "TRBV6-2", NA, NA, "TRBV3-1", NA, "TRBV3-1", 
"TRBV6-4", NA, "TRBV9", "TRBV4-3", "TRBV20-1", "TRAV6", "TRBV6-4", 
"TRBV15", "TRBV20-1", "TRAV30", NA, "TRBV19", NA, "TRBV6-4", 
"TRBV25-1", "TRBV6-2", "TRBV6-1", "TRBV3-1", "TRBV6-6", "TRBV11-2", 
"TRBV27", "TRBV6-6", "TRBV6-4", "TRBV4-2", "TRBV6-6", "TRBV6-1", 
NA, "TRBV6-4")), row.names = c(NA, 100L), class = "data.frame")

df<- data.matrix(df)
xtabs( ~ TRAV+TRBV, data=df)

However, the output of xtabs removes the different levels of TRAV and TRBV columns and instead just provides the numbers. How can I create a matrix and then dataframe as described in the documentation so that I am able to create a chord diagram graph?

Many thanks in advance!

Solution

  • The issue is not xtabs, the issue is that you convert your dataframe to a data.matrix. Instead you can apply xtabs on your dataframe and pass the output to chordDiagram:

    library(circlize)

chordDiagram(xtabs(~ TRAV + TRBV, data = df))

    Or use e.g. aggregate to get an adjacency list as a dataframe and pass that to chordDiagram:

    result <- aggregate(
  rep(1, nrow(df)),
  by = list(TRBV = df$TRBV, TRAV = df$TRAV), FUN = length
)
names(result) <- c("from", "to", "value")

chordDiagram(result)