Hi,

I am facing some difficulties with data transformations of my single cell RNA-Seq data analyzed using DESeq2. This data set is different from typical RNA-Seq experiments.. For example, there is a subset of genes which will be present in one group and totally absent in the other, unlike typical data sets where down regulated genes will still be expressed at lower level. The variation between replicates is also high, and so, we have at least ten replicates for each condition. Even with all these limitations, I am able to get a meaningful result from this analysis.

But the rlog transformation is not optimal for my analysis. I get a warning that more than 10% of the genes have outliers and it suggests doing vst.The vst works for without any warning, but I am still worried if this is optimal or if there is a way to do a better transformation for making heat maps and PCA.

Also, is there a way to extract the normalized values without any transformation? DESeq used to output this, but this function is not in the DESeq2 vignette.

Thanks for your help!

SunilSukumaran

Research Associate

Monell Chemocal Senses Center

Philadelphia

USA

Sorry, I got confused. Is rlg(dds) is logarithm transformation? Then if we just want to normalize the data

without any transformation, counts( dds, normalized=TRUE ) would be the code?Please help me to be clear, I just need either normalization or transformation

rlog and vst give log2-like transformed counts. This is also covered in the vignette in detail.

Thank you,

Sorry, how I can explain the the second principal component between two data sets?? I mean I have applied plotPCA by data from rld, I want to know the reason behind the PCA2 that make two data sets separate not overlapped.

You may want to read some internet references on PCA (there are loads). Briefly PC2 is the dimension which captures the second most amount of variance in the samples, while being

orthogonal(you can think, roughly, "pointing in a totally different direction") to PC1.Thanks a lot

I was plotting PCA with un transformed data

dds=DESeqDataSetFromMatrix(countData = merged,colData = mycols, design = ~ condition)dds.norm <- estimateSizeFactors(dds)plotPCA(dds.norm, intgroup = "condition")Error in (function (classes, fdef, mtable) :unable to find an inherited method for function ‘plotPCA’ for signature ‘"DESeqDataSet"’

How can I plot PCA without transformation please?

I also tried

mat=log2(counts(dds.norm, normalized=TRUE)+epsilon)plotPCA(mat, intgroup = "condition")Error in (function (classes, fdef, mtable) :unable to find an inherited method for function ‘plotPCA’ for signature ‘"matrix"’

hi jivara,

We have extensive documentation. Please read that beforehand. There is a vignette:

https://bioconductor.org/packages/release/bioc/vignettes/DESeq2/inst/doc/DESeq2.html

And a workflow for beginners:

http://master.bioconductor.org/packages/release/workflows/vignettes/rnaseqGene/inst/doc/rnaseqGene.html