Heatmap Error: "X" must be a numeric matrix

[jjanderson4th]

I'm trying to create a heatmap with dendrograms and I keep getting this error: Error in heatmap: "x" must be a numeric matrix. My data table is comparing the deltact for each specific gene (columns) amongst each subjectID that's on the lefthand side.

Here is my csv data table:

	COX2	S100A7	LMNB1	CCL5	CXCL10	OXTR	FLG	KLK6	HMOX1	S100A2
01-0071		10.6
01-0081			2.4						8.57
01-0129										0
01-0169							7.05		4.63
01-0277	5.23	10.8	2.41	3.57						0
01-0281	6.8		3.12	3.38		5.96	7.02
01-0288										0
01-0476					5.17
02-0017					5.13				8.41
02-0023								6.2
02-0056					5.13
02-0069										0
02-0072								3.92
02-0100			2.47
02-0139										0
02-0170					6.98
03-0068										0
03-0076	4.53					6.24
03-0080		6.2			6.26
03-0083	4.71			4.27
03-0091	8.27		4.38	4.11		6.71
03-0116					5.63
03-0131									8.95
03-0150							2.69		9.74
03-0150										0
03-0162	4.12		2.75	4.04	5.85	6.18
03-0221								0
03-0287					6.38			0		0
03-0297										0
03-0355				3.61			7.29
03-0386		11.2
03-0390	4.91			4.91
03-0399										0
04-0027				3.3	6.39
04-0029					4.72
04-0052		9.7
04-0060								5.19
04-0060										0
04-0061							7.47
04-0062							6.98
05-0008	4.76
05-0010				3.47	7.66
05-0013			2.59		5.9
05-0015	5.92			3.5
05-0017				4.51
05-0018			2.39		6.13				4.63
05-0033					5.59
05-0038				3.45	7.15
05-0046	6.75	11.1	2.5			6.32
05-0078			2.91
05-0094									8.32	0
06-0071	4.6	10.7	3.17							0
06-0073						6.07
06-0080	5.36			3.99
06-0094			2.17
06-0119			2.31	3.38	5.91
07-0001						6.34
07-0037			2.85
08-0035	6.75		3.06			7.6
08-0202	4.92	9.3	3.62	4.67		5.93	4.37	5.86		0
08-0215					7.08
08-0217	5.94		2.99
08-0274				3.74
08-0278	6.57	10.2	3.94				6.68	6.44
09-0088		10.7
09-0155				3.39	6.97
09-0177										0
09-0179										0
09-0201				3.35	7.57
09-0246										0
09-0249										0
09-0255				4.75	8.69
10-0036				5.65	7.6
10-0040							7.52		4.63
11-0063				3.69	7.34
12-0047				3.21
13-0009						6.15
13-0038										0
13-0047										0

[AlexisW]

It would help if you provided the code you're using. Trying out something:

# Read in the data I copy-pasted from your post
tab <- read.csv("table.csv")

class(tab)
#> [1] "data.frame"
# take a look at first few rows
head(tab)
#>         X COX2 S100A7 LMNB1 CCL5 CXCL10 OXTR  FLG KLK6 HMOX1 S100A2
#> 1 01-0071   NA   10.6    NA   NA     NA   NA   NA   NA    NA     NA
#> 2 01-0081   NA     NA  2.40   NA     NA   NA   NA   NA  8.57     NA
#> 3 01-0129   NA     NA    NA   NA     NA   NA   NA   NA    NA      0
#> 4 01-0169   NA     NA    NA   NA     NA   NA 7.05   NA  4.63     NA
#> 5 01-0277 5.23   10.8  2.41 3.57     NA   NA   NA   NA    NA      0
#> 6 01-0281 6.80     NA  3.12 3.38     NA 5.96 7.02   NA    NA     NA

# Convert to matrix (make sure to keep the first column as rownames)
mat <- as.matrix(tab[,-1])
rownames(mat) <- tab[,1]

# look at top-left corner
mat[1:3,1:3]
#>         COX2 S100A7 LMNB1
#> 01-0071   NA   10.6    NA
#> 01-0081   NA     NA   2.4
#> 01-0129   NA     NA    NA

class(mat)
#> [1] "matrix" "array"
typeof(mat)
#> [1] "double"


# try plotting it without dendrogram
image(mat)

# this doesn't work because of the many NA
heatmap(mat)
#> Error in hclustfun(distfun(x)): NA/NaN/Inf in foreign function call (arg 10)
pheatmap::pheatmap(mat)
#> Error in hclust(d, method = method): NA/NaN/Inf in foreign function call (arg 10)


mat[is.na(mat)] <- 0
# take a look
mat[1:3,1:3]
#>         COX2 S100A7 LMNB1
#> 01-0071    0   10.6   0.0
#> 01-0081    0    0.0   2.4
#> 01-0129    0    0.0   0.0

# Now we can plot it!
heatmap(mat)

pheatmap::pheatmap(mat)

^{Created on 2022-09-25 by the reprex package (v2.0.1)}

Here I set all the NA to 0 to make distance computation and clustering possible, in your particular case you have to see if it's the right thing to do.