CODEWITHSUNDEEP
rheatmap
Deuterium
Deuterium

Reputation: 101

heatmap2 : Error in axis(1, at = xv, labels = lv) : no locations are finite

I am trying to plot a non symmetrical and non squared matrix using heatmap2. But I am getting a message error :

Error in axis(1, at = xv, labels = lv) : no locations are finite
Calls: heatmap.2 -> axis
Execution halted

In fact, my data are in the matrix are equal to 0, 1 or ? (missing values). For example with this matrix : 

dat_mat
                   C1  C2
P17612|KAPCA_HUMAN  ?   0
P22612|KAPCG_HUMAN  0   1
P22694|KAPCB_HUMAN  1   0
P31751|AKT2_HUMAN   0   0

The expected result should be a heatmap with red color for "0", green for "1" and nothing (blank) for "?". Here is the R script I am using : 

heatmap.2(dat_mat, 
Rowv = FALSE, 
Colv = FALSE, 
dendrogram = "none",
scale = "none",
margins = c(12,24),
key = TRUE,
keysize = 1.0,
col = rainbow(512, start = 1, end = 0.4), 
density.info="density", 
denscol = "black",
xlab="Compounds", 
ylab="Targets", 
main="X Activity Profile",
tracecol = "black")

Upvotes: 2

Views: 3268

Answers (1)

MrFlick
MrFlick

Reputation: 206232

I'm assuming your data is in the following format:

dat_mat<-matrix(sample(c("?","0","1"), 10*2, replace=T), ncol=2)    
colnames(dat_mat)<-c("C1","C2")
rownames(dat_mat)<-letters[1:nrow(dat_mat)]

So you have a character matrix (it's hard to tell from your description). Well, the heatmap is expecting numeric values and it really doesn't like non-standard missing values. So let's remove the "?" and replace them with NA and then convert to numeric

dat_mat[dat_mat=="?"]<-NA
class(dat_mat)<-"numeric"

That's it. You should now be able to plot this as you expect

heatmap.2(dat_mat,
    Rowv=F, Colv=F, dendrogram="none", scale="none",
    margins = c(12,24),
    key = TRUE,
    keysize = 1.0,
    col = rainbow(512, start = 1, end = 0.4), 
    density.info="density", 
    denscol = "black",
    xlab="Compounds", 
    ylab="Targets", 
    main="X Activity Profile",
    tracecol = "black"
)

resulting heatmap

Upvotes: 1

Related Questions