TCGA Clinical Data "Inconsistency"?

Question

[NetunoPoncã]

Hello all,

I downloaded both RNASeq and clinical data from the TCGA website to combine information and perform some data analysis. I am having some problems with the number of clinical data available. For instance, considering: Adrenocortical carcinoma (ACC), the website reports 79 cases for mRNA and 74 clinical cases. The problem is, when I consider all the clinical data I obtained from the website, i.e.,

nationwidechildrens.org_clinical_patient_acc.txt
nationwidechildrens.org_biospecimen_cqcf_acc.txt
nationwidechildrens.org_clinical_cqcf_acc.txt

I end up with a total of 88 unique samples with clinical information (I am interested only in the histological diagnosis for now). What am I doing wrong? Can someone give me a help with this "inconsistency"?

Thanks

[russhh]

having more data than you expected doesn't sound like a problem to me

[Cyriac Kandoth]

Quite often, TCGA receives samples that may fail in any of the many QC steps that lead to good quality RNA-seq. Conversely, there may be some cases that go through the RNA-seq pipelines, but the TCGA never received any related clinical data from the center/hospital that provided the sample... though I don't actually know of any such cases.

Answer 1

[ericfournier3]

Expression data from TCGA may also contain results for matched healthy tissues. You can differentiate healthy tissue samples from tumor samples by looking at the "sample" portion of the barcode. Details here.