TCGA- which files to download for analyzing differentially expressed miRNAs

Question

[ginny]

I have just started working on TCGA data, and I observed that the RNA-seq (HT-Seq counts) files also have the ENSEMBL gene ids for miRNAs, which means that the expression values of miRNA genes are also present in the RNA-seq files.(?)

So then why does TCGA have a separate miRNA quantification dataset (files ending with .mirbase.mirna.quantification)?

I am confused because I plan to find both the differentially expressed genes as well as miRNAs, and don't know which dataset to consider for DESeq2.

Please help! :(

[Barry Digby]

You need to download them separately.

The mirbase.mirna.quantification files are what you want for miRNA DE analysis. You will want to subset the HT-Seq counts too if they contain roughly 50,000 rows (harmonized data) to contain only coding genes ~20,000

[ginny]

Thank you so much! Any idea how can I filter out only the coding genes?

[Barry Digby]

I have code here (https://github.com/BarryDigby/TCGA_Biolinks/blob/master/TCGA_Biolinks.Rmd) that does everything you want: download data, prepare metadata, filtering coding genes, differential expression analysis. It's a good basic template to start with.

It was conducted on TCGA PRAD. Install packages as required, change PRAD to your tissue type of interest and you are good to go.