Hi all,
I'm facing a very annoying error in R while assigning row names to my data matrix. I have some RNA-seq data that I'm considering clustering in R. I'm using gene names as row names for my expression matrix but it keeps reporting that there are duplicate names. Some un-annotated genes have been assigned with some IDs that start with numbers. I don't understand how to deal with this error? Is there a way to work around it? because I cant change the gene names.
EDIT:
gene sample1 sample2 sample3
Mar-01 4.19504 3.9006 4.15683
Mar-02 3.0554 3.4261 3.76675
un_A_2 1.1515 1.2455 0.563484
un_A_3 98.2504 120.341 101.753
ENSGALG00000008227 39.6383 12.8651 38.2281
ENSGALG00000008242 5.71557 7.79314 9.40917
ENSGALG00000008277 24.6231 28.3207 24.9288
CNN3 141.708 134.476 144.514
CNNM1 0.840218 0.963683 0.619086
CNNM2 16.0282 12.1301 12.4665
Many thanks.
2 answers
One way of dealing with this is in R is the function make.names with the option unique=TRUE, see ?make.names.
> nams = c("bl-a","bl-a","bl-a", "foo" )
> df = data.frame(matrix (1:4))
> df
matrix.1.4.
1 1
2 2
3 3
4 4
> rownames(df) = nams
Error in `row.names<-.data.frame`(`*tmp*`, value = value) :
duplicate 'row.names' are not allowed
In addition: Warning message:
non-unique value when setting 'row.names': ‘bl-a’
> rownames(df) = make.names(nams, unique=TRUE)
> df
matrix.1.4.
bl.a 1
bl.a.1 2
bl.a.2 3
foo 4
I am not sure if I understand correctly, but it is best to have unique row names. If you need the original info, you could make a data frame with an additional column containing the unmodified gene names.
It is possible to turn check.names off when constructing the data, but I wouldn't recommend it.
I know this is an old thread, but there doesn't seem to be a good answer. Looking at the example, the names are unique and yet the error is still occurring. I had a similar issue, and the problem was that I had a special character in the name. The special character was handled fine by many r functions but not all. After a lot of troubleshooting, I replaced the special character and this resolved the issue.
make.names should also deal with special characters just fine (by replacing them with "." ) I assumed the example was just an excerpt, and therefor the duplicated names might have been elsewhere.
rownames(x) <- make.names(c("a§", "a€", "aअअअ", "aअअअ"), unique = T)
> x
[,1]
a. 1
a..1 2
aअअअ 3
aअअअ.1 4
Seems to somehow even work with extended unicode, while those characters will not be replaced with dots.
thx,just good work ^_^ ok
I had the same error message when I tried to calculate Spearman correlation coefficient.
> any(duplicated(colnames(data_corr))) [1] FALSE
> any(duplicated(rownames(data_corr))) [1] FALSE
> Corr <- corr.test(data_corr[,-2], use = "pairwise", method="spearman", adjust = "BH", alpha = 0.05)
> Error in `.rowNamesDF<-`(x, value = value) : duplicate 'row.names'
> are not allowed In addition: Warning message: non-unique value when
> setting 'row.names': ‘NA-NA’
The problem was in the class of data frame.
> class(data_corr)
[1] "cast_df" "data.frame"
And resolved by changing class.
class(data_corr) <- "data.frame"
This SO post answer gives more details about "cast_df" class issue:
Log in to answer this question.
Post the sample dataset
I've posted a sample of my data
Gene names - "Mar-01, Mar-02" seems like copy paste from Excel, watch out! http://nsaunders.wordpress.com/2012/10/22/gene-name-errors-and-excel-lessons-not-learned/ http://www.biomedcentral.com/1471-2105/5/80
Hi Diana
Did you get solution for your problem?
Did you read the answer?