I'm new to the world of R, and have been facing this issue and nothing over the internet helped me solve it so am noting it down here.

Thank you for your help.

setwd("~/Desktop/IISc/Melaroma")

library(GEOquery)
my_id <- "GSE4843"
gse <- getGEO(my_id)
length(gse)
## [1] 1
gse <-gse[[1]]
gse

library(WGCNA)
options(stringsAsFactors = FALSE)

expression = as.data.frame(gse[, -c(1)]) 
expression = t(expression)
colnames(expression) = gse$EST
rownames(expression) = names(gse)[-c(1)]

sampleTree = hclust(dist(expression), method = "average")

Error in hclust(dist(expression), method = "average") : 
  NaN dissimilarity value.
In addition: Warning message:
In dist(expression) : NAs introduced by coercion

The proper getter function to get expression data from the ExpressionSet class is exprs() which gives probes as rows and samples as comlumns with colnames and rownames already parsed properly, so you only need that command here:

expression = exprs(gse)

Be sure to follow some sort of tutorial for the hclust part, there are some pitfalls to avoid.