Question: tximport: question concerning 'bias corrected counts without an offset' method
0
5 months ago by
UK,Norwich,University of East Anglia

Hello,

My aim is to use a (kallisto|salmon)/tximport/DESeq2 pipeline using length-corrected gene count estimates.

The tximport vignette provides two methods for importing gene-level counts for DGE (differential gene expression), namely:

Namely, the “original counts and offset” method and the “bias corrected counts without an offset”.

I am fine, with the former, but am having difficulties with the latter.

For the latter method, the tximport vignette advises the following:

"The second method is to use the tximport argument countsFromAbundance="lengthScaledTPM" or "scaledTPM", and then to use the gene-level count matrix txi$counts directly as you would a regular count matrix with these software" For DESeq2, I took this to mean taking txi$counts and then passing this object as the first argument to DESeqDataSetFromMatrix. However, an error relating to the use of non-integer values in txi\$counts results.

So my problem is this: I am not sure if I were to use estimated counts from kallisto/salmon, how I could bias-correct the estimated counts using tximport, and then to use the result as a regular count matrix as suggested by the tximport vigenette, given that matrix values will be non-integer - and unsuitable for use with DESeqDataSetFromMatrix

Apologies in advance, if I have misunderstood something here.

Thanks, Thomas

deseq2 tximport • 126 views
modified 5 months ago by Michael Love24k • written 5 months ago by Thomas Bradley0
Answer: tximport: question concerning 'bias corrected counts without an offset' method
2
5 months ago by
Michael Love24k
United States
Michael Love24k wrote:

Just round() the counts before passing to DESeqDataSetFromMatrix().