https://portal.gdc.cancer.gov/
The Cancer Genome Atlas (TCGA), a landmark cancer genomics program, molecularly characterized over 20,000 primary cancer and matched normal samples spanning 33 cancer types. This joint effort between the National Cancer Institute and the National Human Genome Research Institute began in 2006, bringing together researchers from diverse disciplines and multiple institutions.Over the next dozen years, TCGA generated over 2.5 petabytes of genomic, epigenomic, transcriptomic, and proteomic data. The data, which has already lead to improvements in our ability to diagnose, treat, and prevent cancer, will remain publicly available for anyone in the research community to use.
Q1: How many cases are involved in TCGA database? How many projects?
Q2: How many primary site involved in TCGA database, why the cases size are different amoung various primary sites?
Q3: what are the difference between projects,cases and files? (click "Repository" to see more)
Task1: Find projects about breast cancer RNA-seq cohort.
"project" —— select "breast" in Primary Site and "RNA-seq" in Experimental Strategy
In the end there are 6 projects (1700+ cases).
Task2: Find/Download a WGS mutation information (vcf files) from brain cancer case(age from 20 to 40).
"Repository" —— "Cases" —— select "brain" in Primary Site —— set age in Age at Diagnosis
then "Files" —— select "VCF" in Data Format，click one samples into detail pages and click "Download".
TCGA as a cancer samples data base, you can easily get a opened and small file with "Download" button.
For large file and batch download, you can "Add to Cart" first, then download Manifest file and use GDC Data Transfer Tool in server.
Since most RNA-seq data in TCGA are controlled, if you want to do basic the gene analysis among TCGA samples, online tools GEPIA2 can help you!
http://gepia2.cancer-pku.cn/#index
Task1: Find differential expressed genes in ACC.
"Expression Analysis" —— "Differential Genes" —— select Dataset(cancer name) —— set Log2FC and qvalue cutoff (use default here) —— "List", then you will get the DE genes in ACC.
Task2: Check gene IGF2 expression in tumor samples and normal samples in ACC and BLCA.
"Expression Analysis" —— "Expression DIY" —— "Box Plot" —— add Dataset(cancer name)
Challenge1: Try to do a survival analysis, see if the IGF2 gene expression shows difference survival percentage in ACC.
Challenge2：GEPIA2 is a helpful tool in TCGA RNA-seq basic analysis. Could you please come up with some questions related gene expression in Cancer samples, and try to answer them with GEPIA2?
Above analysis mostly based on gene expression. For mutations in genes, tool cbioportal is helpful.
Task1：check TP53 gene mutations happened in all cancer.
"Quick search" —— select "TP53" in —— "Mutations", you can search keyword to selected cancer type.
Task2：What is the most common hotspot point mutation in KRAS in lung cancer?.
"Query" —— select "Lung" —— select study Lung Adenocarcinoma(TCGA,Nature 2014) —— "Query By Gene"
unselect "Putative CNA from GISTIC" —— type KRAS in Enter Genes —— "Submit Query" —— "Mutations" , G12C
Q1: What is the meaning of "Missense"?
Q2: How many samples are supporting this hostpot point mutation? how to get detail information of those samples?
Task3：use cbioportal to look at difference in survival for lung cancer patients with TP53 and/or KRAS mutations .
"Query" —— select "Lung" —— select study Lung Adenocarcinoma(TCGA,Nature 2014) —— "Query By Gene"
unselect "Putative CNA from GISTIC" —— type KRAS and TP53 in Enter Genes —— "Submit Query" —— "Survival" , Not statistical significant in survival analysis
How about divided into 3 group "TP53 only", "TP53 + KRAS", and "KRAS only".
"overlap" —— select "TP53(75)" and "KRAS(107)" group, we can see there are some overlap samples(25)
group samples：click "82" —— "Creat Group From Seleted Diagram Areas" —— "Submit" (name the other 2 groups in the same way)
select 3 new groups only and click "Survival"
Challenge1:What is the most significantly higher expressed gene in ERCC2 mutant versus WT bladder cancer?