Network construction based on co-ocurrence analysis

The purpose of this app is to construct a network by co-occurence analysis.
The theory behind is there is some connection if two items share the same profile, known as co-occurence
Two files are required for the current version, one is expressin matrix, with rows as entries, column as experiment, the other is grouping information for each enty.
Sample files for download: data matrix file ; meta data file

File format

File format has to be tsv or csv
After uploading your file, please check if they are read in correctly.
Expression data matrix needs to have entry names a row names, while meta data needs to have two columns: sample name and grouping.
If everything is OK, press the button on the right bottom or directly go to next tab to do network analysis.

Upload matrix data(tsv/csv)

Upload meta data(tsv/csv)

Data table pre-process

Q vaule is the presence percentage threshold. For example, 0,75 means 75% of presence across all samples
This function centers and scales numeric matrix. Center means the data(row or column)'s mean is going to be 0. Scale is done after centering. Scale is done by divding each value by their standard deviation. This function is called z-score some elsewhere.

Original Data

Processed Data

Check the matrix table

Check meta data table

Network construction settings

Loading, please wait ...

Contact

Author: Zhibin Ning
Email: ningzhibin@gmail.com
Suggestions and bug report
This shiny app uses visNetwork package for displaying network, for detailed information please refer to visNetwork
more R packages include: shiny, shinydashboard, htmlwidgets, DT, ggplot2, plotly, corrplot, colourpicker, circlize

Change log

V0.3: 20171024. Add moduel for chorddiagram plot and more options for correlation matrix plot
V0.21: 20171017. Add a dedicated module for data processing
V0.2: 20171010. update of the UI, add dedicated buttons for result table downlad
V0.11: 20170910. UI update, a few bugs fix
V0.1: 20170810. functional version onlilne