A web server for epistasis analysis using penalized regression based machine learning methods
EpiML can detect main effect and epistatic effect by using penalized regression based machine learning methods. Interactive figures, graphs, tables, and links to meta-data and external databases will be provided to help retrieve, explore and interpret the identified effects. In addition to these results, users can customize analysis code based on released Jupyter notebooks and pre-configured Docker containers.
Describe job and upload data. Prepare data in requirement format. All user's data will be kept private.
Select a machine learning method. Several popular penalized regression based machine learning methods are built-in.
Automatic processing in background. You don't need to care about the details. Our server will handle all the processing in background.
Check out results.
You can download the results or directly analyze results using interactive visualization.
Interactive tables, figures are provided to help analyze epistatic relationship. EpiML provides friendly interactive visualizations, like circle network, adjacent matrix and force-directed graph, in which users can search and highlight their interested effect.
Links to meta-data and external databases are embedded in visualizations to help interpret the identified effect. UCSC genome browser links are provided for genetic data. miRBase and miR2Disease database are embedded for microRNA data. Multiple resources integration will significantly reduce user time for further analysis.
Jupyter notebook is automatically generated for each job, allowing users to fully customize their epistasis analysis on local computers. All notebooks can be downloaded, modified and rerun based on pre-configured Docker container which contains a fully functional Jupyter server with all the Python and R libraries necessary to run the notebooks.