resem values from TCGA RNAseqV2 data

Question

[Ashwini Patil]

Hi all,

I am working with TCGA RNAseqV2 data and the data consist of rsem values. rsem values have huge variations so I took log2(rsem) and plotted a heatmap. I am interested in seeing expression level. My question is: are rsem values comparable to all sample particularly expression?

Thanks,
Ash

Answer 1

[EagleEye]

Sorry I did not understand your question properly. It will be nice, If you can explain it in more. You can have a look at this.

https://wiki.nci.nih.gov/display/TCGA/RNASeq+Version+2

Calculating Rpkm From Rsem Using Tcga Rnaseqv2 Level3 Data

Note: If the data have huge level of variations. It is always good to log-transform the data for better visualization.

[Ashwini Patil]

I want to know if the RSEM values obtained from TCGA (.rsem.genes.normalized_results) can be used directly for heatmap after log transformation. Should the RSEM values be covered to transcripts per million (TPM)? Should it be also normalized across samples (if so how)? It is difficult to get this information from the TCGA wiki link.

[komal.rathi]

The values are already normalized but for visualization you can use log2(rsemvalue+1) and plot the heatmap.

[Ashwini Patil]

Thanks guys! That helped. My original approach was to remove all values that were < 1. But log2(rsemvalue+1) also sounds reasonable.

For plotting the heatmap, this post helped a lot.