Hello Sir, I downloaded the data, but I am not able to find any normal matched sample, then how should I train my algorithm; could you please suggest some other sites to get miRNA data from; both cancer samples and tumor samples. Thanks in advance!
I downloaded the TCGA dataset for LUAD (Lung Cancer), I find each file with miRNA sequence data like readCount and parts per million value and their are around 123 samples. But where do I find the label, whether this sample is cancerous or not?
Sample Files
TCGA-05-4244-01A-01T-1108-13.hg19.mirbase20.mirna.quantification.txt TCGA-05-4244-01A-01T-1108-13.mirna.quantification.txt TCGA-05-4249-01A-01T-1108-13.hg19.mirbase20.mirna.quantification.txt TCGA-05-4249-01A-01T-1108-13.mirna.quantification.txt
- What is the difference in between isoform and mirna?
- What is the difference in between hg19.mirbase20.mirna and mirna? Should I include both files in my training model?
- Where do I find the label, whether this file data corroborates to a healthy tissue or a cancerous one?
1 answer
To compile clinical phenotypes including case or control, tumor-grade, pathology report etc., you can use one of the secondary databases like CBioPortal or Harnomonizome
- To understand the difference between isoform and mirna, Wikipedia articles are good start
- hg19.mirbase20.mirna = mapped to hg19; mirna = could be the same or annotations using a different refernce database
- You can get the full summary including clinical phenome data here. One of your sample is case (patient with primary tumor) and other is a control; this classification could change based on your case-control criteria)
PS. Not sure from where you downloaded the TCGA data.
Here you go: How To Retrieve Tcga Mirna-Mrna Data
Log in to answer this question.