OncoSG-LUAD mRNA expression (log RNAseq V2 RSEM) uses z-normalized data

[John Lockhart]

As the title says,the mRNA expression (log RNAseq V2 RSEM) plots in the study summary webpage and the data in the data_mrna_seq_v2_rsem.txt file in the study download appear to have been z-normalized across all the samples. The average expression for every gene in the dataset is exactly 0.00.

The data in that file is nearly identical to the data in the data_mrna_seq_v2_rsem_zscores_ref_all_samples.txt file. (A screenshot of the side-by-side comparison is attached)

This is not the case in other datasets like the TCGA-LUAD Firehose.

I've attached some screenshots of example comparisons from the OncoSG and TCGA-LUAD study summary pages to illustrate this point.

Can the correct data please be uploaded for this study?

Thanks,

John Lockhart

[Ramya Madupuri]

Hi John,

Thank you for bringing this to our attention. We confirmed that the RSEM data was normalized and similar to the z-score profile. The data was processed using DESeq2 v.1.16.1, followed by a log2 transformation with a pseudo-count added, and conversion to z-scores. We have temporarily removed the mRNA expression (log RNAseq V2 RSEM) profile from the study and will contact the authors for more information. The z-score profile is still accessible. 

Let us know if you have any further questions.

Thanks,

Ramya

--
You received this message because you are subscribed to the Google Groups "cBioPortal for Cancer Genomics Discussion Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cbioportal+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cbioportal/dafa9290-3a80-4ae6-9683-96cd5ab0c11cn%40googlegroups.com.