Biostar Test Site

This is site is used for testing only. Visit: https://www.biostars.org to ask a question.

Data Discrepency from TCGA-GDC portal and GDC firehose
1
0
Entering edit mode
3.3 years ago
David_emir ▴ 390

Hello All,

I was trying to get Lung cancer samples from TCGA (LUSC & LUAD). I wanted to categorize the data depending upon tumor stage. I have used the following command to download clinical file

library("TCGAbiolinks")

clinical_lusc <- GDCquery_clinic(project = "TCGA-LUSC", type = "clinical")


From GDC-Firehose I was looking for data categorization and they have set of barcodes which they have categorized as "Solid Tissue Normal". For example, one Sample "TCGA-98-8020" is being listed as a Normal sample in gdc-Firehose and the same sample in the Clinical file (got from the TCGAbiolink command) shows as cancerous i.e. the tumor_stage is shown as "stage iiia". As per my understanding, 3rd stage indicates larger cancers or tumors that have grown more deeply into nearby tissue. They may have also spread to lymph nodes but not to other parts of the body. Now I am confused how to interpret this, Is my cancer classification of TCGA samples into Cancerous Vs Normal is correct or I am making something wrong here or both These sites are hosting different data? I am totally confused, Please help.

Thanks a lot, Sincerely, Dave.

TCGA GDC firehose • 1.9k views
0
Entering edit mode

What data are you trying to get? Gene expression, clinical?

0
Entering edit mode

I was planning to Get Differential Gene Expression analysis done on raw count data (RNAseq data) for that i need to group the samples into various levels.

0
Entering edit mode
3.3 years ago
mbk0asis ▴ 620

"TCGA-98-8020" is the ID for participants.

One participant may have both cancer and normal samples.

Two digit code after the participant ID indicates if the sample is cancer or normal.

e.g.)

"TCGA-98-8020-01X-XXX-XXXX-XX" (Cancer)

"TCGA-98-8020-11X-XXX-XXXX-XX" (Normal)


Check out the full sample ID to segregate normals from cancer samples.

0
Entering edit mode

0
Entering edit mode

You are using TCGAbiolinks already, aren't you then also retrieving the full ids for the clinical data?

I assume you're starting with:

query <- GDCquery("yourquery")


Then just type:

getResults(query, cols = c("cases", "tissue.definition"))