Introduction to TRACERx Land

From Array Suite Wiki

Jump to: navigation, search


TRACERx Land for tracing the connection between tumor heterogeneity and tumor progression in NSCLC

The TRACERx (TRAcking Cancer Evolution through therapy (Rx)) lung study plans to transform understanding of non-small cell lung cancer (NSCLC), uncovering mechanisms of cancer evolution by analyzing the intratumor heterogeneity, and tracking tumor evolution from diagnosis through relapse.

TRACERx Land enables exploration of these data by providing access to hundreds of tumor samples with clinical parameters, somatic mutation, and copy number measurements across the exome. The sampling of multiple sites within a given tumor provides a window into how tumors develop and progress, by providing early indications of how new mutations arise.

Land versions

TRACERx Land is built using Human.B38 and OmicsoftGencode.V33.

Data Sources

Tips.pngRNA-seq data are currently not available for this Land

Key Metadata Columns

  • SubjectID
  • TissueCategory/Tissue will not be particularly useful in this Land, as all are lung (use SampleOrigin to filter samples by tissue details)
  • DiseaseCategory/DiseaseState will not be particularly useful in this land, as all are non-small cell adenocarcinoma (use Histology to filter samples by disease details)
  • Histology: Describes the histological classification of each tumor, including invasive adenocarcinoma, squamous cell carcinoma, adenosquamous carcinoma, carcinosarcoma, large cell carcinoma, and large cell neuroendocrine. Use this column to differentiate between invasive and squamous cell carcinomas
  • SampleSource
  • SamplingTime: pretreatment or after treatment, useful for filtering and grouping data chronologically.
  • TumorRegion: In subjects with multiple tumor region samplings, defines R1-R8 for replicates, or LN1/LN2 for lymph node samplings.
  • SampleOrigin: Separates lung tissue samples from lymph node and peripheral blood.
  • WholeGenomeDoubled[WGD][Status]: Whether the whole genome was detected as having doubled in the sample in a clonal or sub-clonal fashion, or not at all.

Key Views

TRACERx contains mutation data, copy number data, and clinical data such as survival information.

Sample Distribution View

Regroup samples on any metadata to get an overview of the available data, and create SampleSets or Custom Queries to subset the data.

For example: regroup based on Histology to reveal the number of samples available for each subtype:


Multi-gene mutation frequency

Search for multiple genes at a time to observe the per-sample mutation frequency of each gene.

As an example, after defining a SampleSet grouping Squamous vs Adenocarcinomas (Histology column), and running "Sample Grouping to Mutation", search for the top 15 genes with differential mutation frequencies.

To remove potential redundancies from multiple mutation samplings of the same subject, use the TumorRegion filter to only look at one sampling (e.g. R1).

Notice that some genes are much more common in squamous adenocarcinoma (TP53, PIKC3A, CDKN2A), but others are more common in invasive (KRAS and AMER3). Change the TumorRegion selection to check whether these patterns hold across sampled tumor regions.


Survival Curves by Sample Grouping

In OmicSoft Lands, you can create Kaplan Meier survival curves comparing any set of samples defined by metadata. New metadata groups can be generated with SampleSets, such as Grouping a SampleSet on the mutation burden across the genome:


From this, you can then plot the survival curve between low mutation burden (<200) or high mutation burden (>200); you could set the cutoffs to any grouping you wish by changing the groups for your SampleSet. It is easy to confirm a key observation of the paper that mutation burden does not play an important prognostic role in these NSCLC patients.