This is a test version of Biostars. For the public version, visit https://www.biostars.org.
DESeq2 vst function error

Hi everyone, I have a dataset with count reads of miRNA-Seq downloaded from TCGA: on the rows I have the miRNA names and on the columns the uuid barcodes(sample names). When I launch the vst function it gets me this error:

dd = DESeqDataSetFromMatrix(d, coldata, design= ~class) 
vst = vst(dd, blind = FALSE)

Error in vst(dd, blind = FALSE) : 
  less than 'nsub' rows with mean normalized count > 5, 
  it is recommended to use varianceStabilizingTransformation directly

This is the first attempt with this package, any suggestion? Thank you so much

deseq2 rna-seq r normalization
• 8,176 views • link
updated by DareDevil DareDevil • written by lenC_biotecLover • lenC_biotecLover

2 answers

Your dataset is small (< 1000 variables) and / or has a low number of expressed genes. Either way, the problem can be mitigated by using varianceStabilizingTransformation(), or just re-use vst() and set a lower value for nsub.

Type ?vst to see what is happening with the nsub parameter.

Kevin

• 0 views • link
written by Kevin Blighe Kevin Blighe

I set a lower number, and it works, thank you! My dataset has 1800 miRNAs. This does mean that only few of them are considered for the calculation? How does the algorithm operates in this cases?

• 0 views • link
written by lenC_biotecLover • lenC_biotecLover

You can also use assign a value to nsub:

nsb = sum(rowMeans(counts(dds, normalized=TRUE)) > 5)
vst(dds, blind = TRUE, nsub = nsb)
• 0 views • link
written by DareDevil DareDevil

Log in to answer this question.

More posts like this