/tcga-shiny

Shiny app to explore data from The Cancer Genome Atlas

Primary LanguageR

tcga-shiny

This repository contains an R Shiny app to explore data from The Cancer Genome Atlas (TCGA). The data used has been processed with scripts from https://github.com/arnijohnsen/tcga-analysis

Installation

Shinyapps.io

The app is available at https://arnijohnsen.shinyapps.io/tcga-shiny

Running locally

To run the app locally, use the following commands within R

# install.packages(c("shiny", "ggvis"))
library(shiny)
runGitHub("tcga-shiny", "arnijohnsen")

You may need to install the shiny and ggvis packages before running the app. If so, run the commented line in the commands above.

Cloning repository and running locally

On *nix systems, you can clone this repository, run the app and make modifications with the following commands

git clone https://github.com/arnijohnsen/tcga-shiny
R
> library(shiny)
> runApp("tcga-shiny")

User guide

Selecting genes, probes and axis variables

Begin by selecing or searching for a gene by it name, in the "Search for gene" textbox. Once a gene has been selected, you can select a methylation probe linked to the selected genes, in "Search for probe". Each probe is annotated based on its position in the gene:

  • (promoter) indicates probes in TSS200 or 5'UTR regions
  • (body ...) indicates probes in gene body*
    • (body_island) indicates probes in gene body inside cpg islands
    • (body_shore) indicates probes in gene body on cpg island shores
    • (body_none) indicates probes in gene body outside cpg islands
  • (enhancer) indicates probes linked to enhancer regions
  • (undefined) indicates probes which fall in to no other category

If a probe falls in to multiple categories, the categorical hierarchy is: promoter, enhancer, body, undefined

Three numerical variables are available for plotting:

  • Copy number variation
  • Gene expression from RNA Sequencing
  • Methylation

If methylation is selected as a plotting variable, a methylation probe must also be selected in "Search for probe".

Color points by categorical variable

The scatterplot can be colored by 3 types of categorical variables, chosen in the first "Select variable to color points" textbox:

  • Somatic mutations in the gene
  • Tumor subtype
  • Clinical information

For tumor subtypes and clinical information, a secondary classification is chosen from the second "Select variable to color points" textbox.

For tumor subtypes, three classifications are available:

  • PAM50 subtypes based on microarray data
  • PAM50 subtypes based on RNA Sequencing
  • iC10 subtypes

For clinical information, nine classifications are available:

  • Age
  • Menopause status
  • Tumor status (is the participant tumor free)
  • Vital status of participant
  • Tumor stage
  • ER (Estrogen receptor) status by IHC (immunohistochemistry)
  • PR (Progestrone receptor) status by IHC
  • HER2 (Human epidermal growth factor receptor 2) status by IHC
  • Histological type of tumor

Change point size and opacity

Sliders can be used to change point size and opacity, as parts of some plots can become cluttered and difficult to see individual points.

Change plot dimensions

The plot width is automatically set as the browser width. Resize your browser window to change plot width.

The plot height is set by a slider, which defaults to 800 px.

Hover tooltips

Once a plot has been generated, you can hover over individual points to display the following information:

  • Participant barcode
  • x-value for the point
  • y-value for the point
  • Value of categorical variable used to color points

Save plots

Once a plot has been generated, you can save the plot as either .svg or .png file. Click the cogwheel at top right of the plot to choose between SVG or Canvas, and click "Download SVG" or "Download PNG" to save the plot.