SparseData Cluster
Objective
The goal of this project is to build a beautiful parser of data that can interpret matrix data (with a specific use-case being gene expression matrices) and construct basic interactive plots for data exploration and preliminary analyses.
Get Started
Use the online version of SparseData Cluster. See Installation for details on installing locally.
Functionality
- Upload : Upload your own flat files (comma, tab, or semi-colon delimited) for analysis.
- Cluster : Pair-wise correlation is computed between observations (by default, rows of matrix input) and displayed as a heatmap. A summary of the matrix is also given as plain text.
- Rank : Choose 2 observations to view an interactive table of the differences for each feature. Note that when data is log2 transformed during Upload, these will correspond to log fold changes.
Details
Installation
Dependencies
- This App depends on installation of the following R packages:
shiny
(version >= 0.12.1),shinydashboard
,shinyapps
,markdown
,gplots
,RColorBrewer
.
To Run:
Open app.R and run the code in an interactive R session in the same directory
Organization
The application is organized into separate files as follows:
app.R
: The top-level application that sources the rest of the necessary files to build the app and calls theshinyApp
functionglobal.R
: Globally needed packages and global variables to share data across multiple embedded appsheader.R
: constructs the header barsidebar.R
: constructs the sidebar; specific pages are delineated here via thetabName
function, and are similarly defined inbody.R
body.R
: page-level construction of eachtabName
specified insidebar.R
server.R
: provides interactivity and backend calculations
Acknowledgements
Stefan Avey constructed the underlying base, and Rob Amezquita applied a slick coat of paint on it.