Correlation analysis between gene expression & tumor purity

Question

[newbio17]

What I have seen previously in papers is that the authors find genes that are highly correlated with tumor purity.

• Paper #1
• Paper #2

I'm mainly confused in Paper #1, when author says:

"Spearman correlations were calculated between their expression and tumor purity to generate a ranked list of genes."

If you have a matrix consisting of 50 samples where they are separated into groups of 10 according to the tumor type (5 groups) and 20,000 rows for genes, how do you perform correlation analysis with 50 tumor purity values for the 50 samples to identify the genes that are highly correlated?

Answer 1

[Devon Ryan]

They just computed the correlation between each row of gene expression values (50 values) and the 50 purity values and then creating a ranked list of the results. In R you would do that with cor() and apply().

[newbio17]

Thank you. Your posts are always helpful.