Error in DESeqDataSetFromMatrix Function in DESeq Library

Question

[davidenoma]

I am trying to create a DESeq dataset object from a dataset in GEO as follows:

deseq2_142731 <- DESeqDataSetFromMatrix(countData = GSE142731[,2:ncol(GSE142731)],colData = labels_gse142731,design = ~V1)

However, I get an error:

Error in DESeqDataSet(se, design = design, ignoreRank) : some values in assay are not integers

Both the colData and countData variables are from GEO, they have the same dimension. Please, any ideas on how to handle this?

Answer 1

[ATpoint]

some values in assay are not integers

DESeq2 takes raw counts (integers) as input. So if you are sure that this table you downloaded are in fact raw data then you can just round() them to make integers. This is in line with the advise from the DESeq2 author, e.g. here https://support.bioconductor.org/p/88763/#88764

Quote:

I'm not worried about any loss of precision for inference of log fold change in this rounding, because fractions of counts are tiny compared to the sampling and biological variation on counts in RNA-seq.

[swbarnes2]

However, if you have non integer data, you should stop and think and make sure that you have counts, and not TPM or some other normalized values.