GTEX vs TCGA rnaseq data

[Gabriela]

Hi,

I'm working with the TcgaTargetGtex_RSEM_Hugo_norm_count dataset. Can a compare the expression values of GTEX and TCGA samples? Are these expression values upper quartile normalized?

Many thanks!

Gabriela Kinker

[Jing Zhu]

>Are these expression values upper quartile normalized?

N

​o. They are not.

Jing​

--
You received this message because you are subscribed to the Google Groups "UCSC Xena and Cancer Genomics Browser" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ucsc-cancer-genomics-browser+unsub...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Gabriela]

Thank you for your answer,  Jing.

Can I compare the expression values of these GTEX and TCGA samples?

Gabriela Kinker

[Jing Zhu]

Without data standardization, i.e. upper quartile normalized, there will definitely be more noise introduced. Ideally, you want to preform data standardization. 

>Can I compare the expression values of these GTEX and TCGA samples?

​Wearing my hat of xena browser support, I​

 can't really

​make that decision for you. Sorry. 

We are aware of the data standardization needs, and it is on our radar to address it on the browser site without changing the underlying data (i.e. the underlying data you download for whole-genome analysis will still be un-standardized).

Jing

To unsubscribe from this group and stop receiving emails from it, send an email to ucsc-cancer-genomics-browser+unsubs...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Carter Barger]

Hi Jing, 

As follow-up question to GTEX and TCGA comparisons. Can I use the RNA-seq data to compare expression for a single gene between GTEX and TCGA cohorts (mean, median comparisons)? For example, GTEX has normal samples whereas some TCGA cohorts do not have normal samples for expression based comparisons. 

Thanks, 
Carter

On Monday, May 8, 2017 at 1:40:25 PM UTC-5, Jing Zhu wrote:

Without data standardization, i.e. upper quartile normalized, there will definitely be more noise introduced. Ideally, you want to preform data standardization. 

>Can I compare the expression values of these GTEX and TCGA samples?

​Wearing my hat of xena browser support, I​

 can't really

​make that decision for you. Sorry. 

We are aware of the data standardization needs, and it is on our radar to address it on the browser site without changing the underlying data (i.e. the underlying data you download for whole-genome analysis will still be un-standardized).

Jing

On Mon, May 8, 2017 at 10:51 AM, Gabriela

Thank you for your answer,  Jing.

Can I compare the expression values of these GTEX and TCGA samples?

Gabriela Kinker

Em segunda-feira, 8 de maio de 2017 14:36:01 UTC-3, Jing Zhu escreveu:

>Are these expression values upper quartile normalized?

N

​o. They are not.

Jing​

On Sun, May 7, 2017 at 2:17 PM, Gabriela

Hi,

[Mary Goldman]

Hi Carter,

I'm sorry but it's the same response as last time: we can't make that decision for you. Some people have made that comparison and others have not. You'll have to decide for yourself.

Best,
Mary
-------------
Mary Goldman
UCSC Xena Browser
http://xena.ucsc.edu/

[carter...@gmail.com]

Thanks! That's fine. I understand. Could you please explain the technical issues for making these comparisons so I can better understand and make my own decision?

Sent from my iPhone

[Mary Goldman]

Hi Carter,

Mainly people are concerned about batch effects as RNAseq data is very prone to them. There are many primers out there on batch effects. This paper is one of many that may help you decide: http://www.nature.com/nrg/journal/v11/n10/full/nrg2825.html

[Carter Barger]

Thanks Mary!

To unsubscribe from this group and stop receiving emails from it, send an email to ucsc-cancer-genomics-browser+unsub...@googlegroups.com.