From known genes to new hypotheses
Suppose you have a set of genes associated with a disease or biological process. Which other genes are connected to them, and does the answer change when you combine several kinds of evidence? RandomWalkRestartMH provides a way to investigate these questions using network proximity.
How random walk with restart works
The analysis starts from selected seed nodes. A walker explores their connections and periodically returns to the seeds. The resulting scores describe how closely other nodes are connected to this starting point. These scores support candidate prioritisation and can be considered alongside experimental evidence and the coverage of the underlying networks.
Keeping different types of evidence connected
- Multiplex networks represent the same genes or proteins in several interaction layers.
- Heterogeneous networks connect different kinds of entities, such as genes and diseases.
- Combined networks allow an analysis to move between interaction layers and entity types.
Get started in R
The maintained package is available from GitHub. Install it with pak, then follow the worked examples in the vignette.
# Install pak first if needed:
install.packages("pak")
pak::pkg_install("alberto-valdeolivas/RandomWalkRestartMH")
Read the worked network-analysis tutorial ↗
The research behind the package
The method is described in Random walk with restart on multiplex and heterogeneous biological networks, by Valdeolivas and collaborators, published in Bioinformatics in the 2019 issue and online in 2018.
This overview summarises the package and its research context. Detailed usage, examples and citation information are maintained in the linked package documentation.