Entering edit mode
8 weeks ago
Znow
▴
20
I would like to compare RNA-seq data from TCGA BRCA with GEO microarray. I am going to make a risk score for predicitng prognosis bases on the coefficients from a lasso-cox model multiplied by the gene expression. Most of the studies available in pubmed trains their models in TCGA RNA-seq data and tests it in GEO microarray datasets. However, they use different measures, how is it possible to compare them? Can someone please help?
Please take a look at these past threads on this topic:
Comparing microarray to RNA-seq data
Combining Microarray and RNAseq data
TCGA microarray comparison with RNA-seq
How to scale data between two different platforms: RNA Seq and Illumina
As you mentioned, you can use TCGA as the discovery set, and use GEO as validation. The data is never mixed together during analysis, and you are simply compare the findings.
Don't mix and compare the data directly.