8.2 Detailed Examples using Breast Cancer Data

Last modified by Erik S on 2019/05/29 16:29

A few examples of the types of explorations possible with Regulome Explorer: All Pairs are described below in the context of Breast Cancer analysis.  These examples were part of the TCGA breast cancer marker paper (Nature 490, 61-70, 2012).

Load Data

You can generate these views by clicking on this link:

explorer.cancerregulome.org/all_pairs/?dataset=brca_manuscript_rerun_nov12d_pw

Exploring significant associations between molecular features

Filtering for significant associations between microRNA features and gene expression (mRNA) features in which there is a negative correlation results in the circular view shown in the figure below, in which each arc indicates an association between a microRNA and a gene. Hovering over a single arc displays an Edge Card, and clicking on the arc produces a Details Window of the underlying data. The majority of the arcs are due to microRNAs hsa-mir-17/18a at chromosome 13q31, hsa-mir-190b at 1q21, and hsa-mir-210 at 11p15.

https://lh5.googleusercontent.com/NRMuNgsQ162cOsnPsEL_3AiYbRB5SJSN2hKKc7TvxsCMruTOh9Lk4GetF8BfXbngVOZ59efFeJr3JAJDBe1ltfLK_yMN1Wmj6OlQRnvR_MYOwUWsfVY


To generate this view, set the filter to the parameters shown in this filter window:

Regulome Explorer All Pairs - brca_mirn_gexp.png

While microRNAs can affect distal genes, one would expect copy-number aberrations to primarily affect the expression of proximal genes, and this can be seen in the figure below in which the most significant associations between copy-number features (orange) and gene expression features (blue) are nearly always proximal and are concentrated in hot spots on chromosome arms 1q (e.g., PARP1) and 5q (e.g., REEP5 and IL6ST) and chromosomes 8 (e.g., BRF2 and RAD21), 16 (e.g., CENPN) and 17 (e.g., ERBB2 and MIEN1).

https://lh5.googleusercontent.com/buo59Aai7yPgUQ2_GNmhobnInMrkaSADtC0AbIbhUvaQhfxXX8W2T2Ij132LftdnUm2eTGc4Vf4A9IHlj8D7wwKoN7LjegKBoYjJvMkOwDYHRhhyFyE

To generate this view, set the filter to the parameters shown in this filter window:

Regulome Explorer All Pairs - brca_scna_gexp.png

Exploring significant associations with a tumor subtype

Associations between molecular or other features and categorical features representing clusters in the population can also be explored. In this case, statistically significant associations between molecular features (with genomic coordinates) and a categorical feature are shown as dots on a circular graph with a radial axis representing correlation coefficient or -log10 (p-value).

For example, CpG dinucleotides that are significantly differentially methylated in the hyper-methylated cluster shown in main Figure 2 of the TCGA breast cancer paper can be identified by filtering for features associated with a binary feature that indicates membership in that cluster, as shown in the figure below.

https://lh4.googleusercontent.com/1NsUMEDWLqez0Bwdtt2h0H7ToSlvGZQ5V0ykmr7LRgvnxuhQp9pUgaxDpNa1fMnJMuOGZpnp9yHPEYZYFEXURcMNuluckF2o6uhDJzqYAjWhcd6xj00

Proteins that are significantly differentially expressed between the RPPA-inferred Reactive I and Reactive II groups shown in main Figure 3 of the TCGA breast cancer paper can be identified as shown in the figure below.


https://lh5.googleusercontent.com/izeR7sNtQTQGSPkSTGLDuc4KtvmUC5fBv999bVvddYsG-9cEaLxQnjhVFq7159EUct60yljA5QxIsvPvwDl-_RQ9bK3jJPuGUR70l4jc-mq46lIeKD4

Tags:
Created by Erik S on 2019/05/29 16:29

Need help?

If you need help with XWiki you can contact: