TCGA data rsem

[arash nabbi]

Dear All,

I am very new into this. I am analyzing TCGA data to see the differential expression of a specific gene across samples. I downloaded both rsem.genes.results and rsem.normalized.results. There are many different opinions on this. Which one can I use to do the analysis? There is raw_count and scaled_estimate. From what I read, raw_count can be used as input in edgeR. But scaled_estimate is like transcripts per million and therefore can be used directly with no transformation. And also, I can not use normalized_count to compare across samples, but again I see many papers do a log2 transformation on normalized_count and use that....

I am pretty confused.
Please help
Thanks
Arash

[Ning Leng]

Hi Arash,

I believe you'll need to input the raw count into edgeR. The log2 transformed normalized expression are more widely used for visualization purposes.

Hope this is helpful.

Best,

Ning

--
RSEM website: http://deweylab.biostat.wisc.edu/rsem/
---
You received this message because you are subscribed to the Google Groups "RSEM Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rsem-users+...@googlegroups.com.
To post to this group, send email to rsem-...@googlegroups.com.
Visit this group at http://groups.google.com/group/rsem-users.

--

Ning Leng, Ph.D
Computational Biologist / Biostatistician
Regenerative Biology Laboratory
Morgridge Institute for Research

[Arash Nabbi]

Thanks Ning,

The problem is when I do log2 there is very little difference across samples (7.8 vs 8.1). Is this something I should expect after transformation? When I do z-score (with reference population being normal tissue), the difference is pretty significant, but it is hard to graph it. So I do not know what I should report

Arash

You received this message because you are subscribed to a topic in the Google Groups "RSEM Users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/rsem-users/xh4HtI87glk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to rsem-users+...@googlegroups.com.