influential influential logo

Network influence analysis, centrality assessment, and experimental feature prioritization in R

R-CMD-check CRAN Version Total CRAN Downloads Monthly CRAN Downloads Integrated Value of Influence: IVI SIR-based Influence Ranking: SIRIR Experimental data-based Integrative Ranking: ExIR

Overview

influential is an R package for identifying influential nodes in networks and for classifying and prioritizing candidate features from experimental data. It brings together association analysis, network reconstruction, centrality assessment, influence ranking, visualization, simulation, and experimental feature prioritization in a single toolkit.

Two complementary workflows sit at the center of the package:

The package also provides Hubness score for local network power, Spreading score for information-spreading potential, SIRIR for simulation-based influence ranking, computational knockout/up-regulation, fast correlation analysis, network reconstruction utilities, centrality measures, and centrality-based network visualization.

At a glance

Capability What influential provides
Association analysis Fast Pearson/Spearman correlation analysis with optional mutual rank, p-values, and adjusted p-values
Network reconstruction Construction of igraph networks from data frames, adjacency matrices, incidence matrices, and SIF files
Centrality analysis Local, semi-local, and global centrality measures together with association assessment
Network influence IVI, Hubness score, Spreading score, and SIRIR
Experimental prioritization ExIR-based classification and ranking of drivers, biomarkers, and mediators from omics data
Perturbation & visualization In silico knockout/up-regulation and centrality-based network visualization
Interactive analysis Browser-based and locally launchable Shiny interfaces for IVI and ExIR
Resource Link
✨ Feature Explorer Explore the features and capabilities of influential
📦 CRAN cran.r-project.org/package=influential
📖 Full vignette Introduction to influential
💻 GitHub github.com/asalavaty/influential
🌐 Interactive web portal Influential Software Package
🐞 Issues & feature requests GitHub issue tracker

Installation

Install the current CRAN release:

install.packages("influential")

Or install the development version from GitHub:

# install.packages("remotes")
remotes::install_github(
  "asalavaty/influential",
  build_vignettes = TRUE
)

Then load the package:

library(influential)

Core methods

Integrated Value of Influence (IVI)

IVI integrates complementary local, semi-local, and global centrality dimensions to identify influential nodes within a network.

For the methodological details, see:

Integrated Value of Influence: An Integrative Method for the Identification of the Most Influential Nodes within Networks

Experimental data-based Integrative Ranking (ExIR)

ExIR prioritizes candidate features directly from experimental omics data by integrating multiple levels of evidence with network reconstruction and influence ranking. Depending on the input data and analysis settings, ExIR can identify drivers, biomarkers, DE-mediators, and nonDE-mediators.

For the methodological details, see:

ExIR enables prioritizing driver and biomarker genes from omics data in a reference free manner

Documentation

A comprehensive introduction to influential and its functions is available in the package vignette:

Read the influential vignette

You can also browse installed vignettes directly from R:

browseVignettes("influential")

Shiny apps

The package provides interactive interfaces for IVI and ExIR through the Influential Software Package web portal.

IVI Shiny app

The IVI Shiny App supports calculation of IVI values and IVI-based network visualization.

You can also launch it locally:

influential::runShinyApp("IVI")

ExIR Shiny app

The ExIR Shiny App supports ExIR analysis, result visualization, and downstream exploration.

You can also launch it locally:

influential::runShinyApp("ExIR")

How to cite influential

If you use influential, please cite the publication associated with the method(s) used in your analysis.

For the Experimental data-based Integrative Ranking (ExIR) model:

For the Integrated Value of Influence (IVI) and network influence analysis:

Package citation information is also available from R:

citation("influential")

Author

The influential package was developed by Adrian Salavaty.

Advisors

Contributing and support

Bug reports, feature requests, documentation suggestions, and other contributions are welcome.

Please use the influential GitHub issues tracker to report problems or suggest enhancements.