Comparing gene expression in cancer vs. control

Question

[Dirk]

I've obtained some RNASeq data thats been TPM corrected for a cancer of interest from the CCLE (primarily interested in AML). How does one traditionally compare the TPM expression of an arbitrary gene in a cancer vs control? The CCLE doesn't seem to provide any control/baseline data so it's hard to know if your gene of interest is really at all overexpressed from the CCLE standpoint.

[newbio17]

I'm not sure how helpful this is, but you can try going through the cell lines on atcc.org. AML is a little tricky, but for other cancer types, they have corresponding normal cell lines per tissue type available here. Once you get the names of the cell lines, go to either GEO or SRA to search for the data. You will probably will need to quantify the gene expression yourself.

Answer 1

[Kevin Blighe]

In such a scenario, we could go by Z-scores, whereby Z > 1.96 or Z < -1.96 are statistically significantly up- and down-regulated, respectively, at 5% alpha. Prior to Z-score transformation, in this case, it may make more sense to only have AML cell-lines in your dataset - if you have all other cell-lines, then the result and meaning / interpretation will change.

Kevin