OncoLand Case Study - Clinical Variables
From Array Suite Wiki
Explore Clinical Variables to identify novel associations
TCGA data come with extensive clinical metadata. TCGALand contains these as extensively curated tables associated with each sample. TCGALand also provides functions to explore and visualize these clinical variables, and how they correlate with other data.
Explore Clinical Associations with MDM2 expression
To explore how a gene's expression level (or mutation, copy number, etc) correlates with clinical metadata, you can create a "Custom Query". In this example, Mdm2 RNA-seq expression is analyzed for correlations with TCGALand Clinical Metadata, revealing association with expression status of several critical receptors.
Identify associations between clinical variables
One powerful feature of OncoLand is Group Association, where the entire set of Clinical Metadata are analyzed for statistically significant association with the specified grouping. By default, grouping is by tumor, but you can group by metadata, SampleSets, or Custom Queries.
Explore how gene mutations correlate with clinical variables
To explore how a mutations in a gene correlates with clinical metadata, you can create a "Custom Query". In this example, tp53 mutation status is analyzed for correlations with TCGALand Clinical Metadata in lung cancer, finding associations with tobacco smoking metrics.
Find correlations between a gene's mutation and cancer subtype gene expression signatures
A recent paper identified mutations in four genes that correlate with "Gene Signatures" for subtypes of glioblastoma. By simply querying TCGALand for the mutation status of these four genes, the "Group Association" module reveals these associations, which can be further explored in OncoLand Views.