Hello! I'm interested in plotting the expression values of my samples for a certain gene. I'm a little confused on how to do this. In my pipeline, I used kallisto to estimate transcript abundance and then used tximport to analyze expression within DESeq2. So far, I generated some plots using the TPM counts from tximport (which I had given to DESeqDataSetFromMatrix function as the countData argument).

My confusion is that TPM is considered a normalized metric and I have seen plots of TPM in papers before (typically as log TPM). I was re-reading the DESeq2 manual and found a post on here that recommended using counts(dds,normalized=TRUE).

I wanted to confirm that the difference between TPM counts from tximport and the counts(dds,normalized=TRUE) function in DESeq2 is that TPM just corrects for transcript length, and counts(dds,normalized=TRUE) corrects for transcript length and library size?

In that case, if I wanted to calculate mean TPM of a gene and visualize expression (log TPM) via box plot, that I should use the counts(dds,normalized=TRUE) instead of TPM?

You can plot TPM if you like. This is a measure of abundance.

You can access TPM if you use tximport or tximeta to bring in the data:

# this is a vector of TPM:
assay(dds, "abundance")["geneX",]
# you can plot this vector along with sample covariates e.g.
dds$condition