Having trouble using the SAFE clustering method

Question

[Chynoki]

Hi, I would like to use SAFE clustering on my single-cell dataset, but it seems the scripts are bugged. I tried to run the example provided on SAFE's website, but I'm getting this error.

Error in individual_clustering(inputTags = expr.mat, SEED = 123) :
(list) object cannot be coerced to type 'double' In addition: Warning message: 'newSCESet' is deprecated. Use 'SingleCellExperiment' instead. See help("Deprecated")

Here is the code I ran in order to obtain this error:

expr.mat <- read.csv("/home/snin/workspace/SAFE_clustering/expression_matrix.csv", header = T)
rownames(expr.mat) <- expr.mat[,1]
expr.mat <- expr.mat[,-1]
cluster.result <- individual_clustering(inputTags=expr.mat, SEED=123)

Can you maybe advise me to resolve this issue?

Have a nice day.

Chynoki

Answer 1

[Kevin Blighe]

Try

cluster.result <- individual_clustering(data.matrix(expr.mat), SEED=123)

If that does not work, then check your object to ensure that there are no characters in it. It has to be numeric. In some rare situations, a trailing space may appear after a number, in which case R will become confused. You can detect these by just converting the entire thing to text with:

as.character(expr.mat)

Answer 2

[Chynoki]

Hi, thanks for your answer.

I also contacted the dev of the tools and they told me the same. I used SAFEv1.0 instead of SAFEv2.0. Use SAFEv2.0 resolve my problem. It was the SCE package that were updated to v1.6 and SAFEv1.0 use SCEv1.3.

Now I have an other problem when running the tutorial.

This line:

ARI.SC3 = adjustedRandIndex(cluster[1,], celltype.info$cluster)

Give me the following error:

Error in as.vector(x) : object 'cluster' not found

I understand the error, but in the tutorial they do not instantiate a 'cluster' object.

I'm also waiting for an answer from the dev.