A Roadmap to The Molecular Human - Linking Multiomics with Population Traits and Diabetes Subtypes
This is the GitHub page for the comics multiomics web-server.
Comics is a shiny interface to the The Molecular Human and provides a visualization of over 34,000 associations between over 8,700 multiomics traits and disease endpoints.
The underlying networks are derived from 18 technically diverse deep molecular phenotyping (omics-)platforms analyzing urine, blood, and saliva samples from up to 391 participants of the multi-ethnic diabetes case-control study QMDiab.
The raw data used to generate the Comics network is freely available on FigShare.
The links between the multiomics traits include partial correlations between traits from the same platform (GGMs), mutual best hits for pairwise all-against-all correlations between platforms, genome-wide, epigenome-wide and transcriptome-wide multiomics and disease associations, and associations with QMDiab clinical endpoints.
Details of the comics server and its underlying multiomics networks can be found on the comics homepage.
The comics web server is available following this link: http://comics.metabolomix.com.
You need to download and install docker as described here.
Then pull the docker image ghcr.io/karstensuhre/comics using the docker GUI or run the following command line:
docker run --detach --name comics -p 8080:3838 ghcr.io/karstensuhre/comics
Navigate to your browser and open the following page: http://localhost:8080.
All files required to run comics locally using RStudio are in the ./assets directory of this GitHub repository. Start-up rstudio with the ./assets as working directory and then launch app.R as a shiny app.
Note: you need to have the following libraries installed in RStudio:
install.packages("readxl")
install.packages("tidyverse")
install.packages("visNetwork")
install.packages("shiny")
install.packages("shinydashboard")
install.packages("DT")
Here is a screenshot of the comics server:
The comics server is a convenient interface to The Molecular Human. It allows to navigate the neighborhoods of individual multiomics and disease traits. However, more complex network analyses require more spohisticated tools. For this purpose we provide a cytoscape version of the full network with matching shapes and colors.
Download and open this file using cytoscape.
Here is a screenshot of the The Molecular Human opened in cytoscape:

