Network-Based Integration of Multi-Omics Data Using Sparse CCA

Provides an end-to-end workflow for integrative analysis of two omics layers using sparse canonical correlation analysis (sCCA), including sample alignment, feature selection, network edge construction, and visualization of gene-metabolite relationships. The underlying methods are based on penalized matrix decomposition and sparse CCA (Witten, Tibshirani and Hastie (2009) ), with design principles inspired by multivariate integrative frameworks such as mixOmics (Rohart et al. (2017) ).

• OmicNetR
  • Installation
  • Quick start
  • Figures
  • Included example dataset
  • Main functions
  • Code of conduct

OmicNetR

OmicNetR is an R package for integrative multi-omics analysis using Sparse Canonical Correlation Analysis (sCCA). It supports an end-to-end workflow for aligning multi-omics datasets, fitting an sCCA model, creating a gene–metabolite bipartite network, and generating summary visualizations.

Installation

Development version (GitHub)

install.packages("devtools")
devtools::install_github("ppchaudhary/OmicNetR")

library(OmicNetR)

Quick start

set.seed(123)

omics_data <- generate_dummy_omics(
  n_samples = 60,
  n_genes = 800,
  n_metabolites = 150,
  n_linked = 20
)

X <- omics_data$X
Y <- omics_data$Y

aligned <- align_omics(X, Y)

## Successfully aligned 60 matching samples.

scca_model <- omic_scca(
  X = aligned$X,
  Y = aligned$Y,
  n_components = 2,
  penalty_X = 0.70,
  penalty_Y = 0.70
)

## Model Optimization: Keeping 240 genes and 45 metabolites.

net_data <- scca_to_network(
  scca_model,
  comp_select = 1,
  weight_threshold = 0.01
)

# Keep top edges for a cleaner plot
net_data <- net_data[order(abs(net_data$Weight_Product), decreasing = TRUE), ]
net_data <- head(net_data, 50)

Figures

The script tools/render_readme.R will (1) generate figures into man/figures/ and (2) render README.md. If you already generated figures, the blocks below will display them in GitHub.

Bipartite gene–metabolite network

Feature importance circle plot

Gene–metabolite correlation heatmap

Included example dataset

data(omics_example)
str(omics_example, max.level = 1)

## List of 3
##  $ X       : num [1:50, 1:200] -1.6301 0.0816 -0.0224 -0.588 0.5384 ...
##   ..- attr(*, "dimnames")=List of 2
##   ..- attr(*, "scaled:center")= Named num [1:200] -0.0411 0.3266 -0.1829 0.015 -0.3237 ...
##   .. ..- attr(*, "names")= chr [1:200] "Gene_1" "Gene_2" "Gene_3" "Gene_4" ...
##   ..- attr(*, "scaled:scale")= Named num [1:200] 2.03 1.68 2.28 1.91 2.14 ...
##   .. ..- attr(*, "names")= chr [1:200] "Gene_1" "Gene_2" "Gene_3" "Gene_4" ...
##  $ Y       : num [1:50, 1:50] 0.279 -0.754 -1.271 1.383 0.084 ...
##   ..- attr(*, "dimnames")=List of 2
##   ..- attr(*, "scaled:center")= Named num [1:50] -0.1551 -0.1521 0.2014 -0.0142 0.2227 ...
##   .. ..- attr(*, "names")= chr [1:50] "Met_1" "Met_2" "Met_3" "Met_4" ...
##   ..- attr(*, "scaled:scale")= Named num [1:50] 2.39 2.27 2.26 2.38 2.24 ...
##   .. ..- attr(*, "names")= chr [1:50] "Met_1" "Met_2" "Met_3" "Met_4" ...
##  $ metadata:'data.frame':    50 obs. of  3 variables:

Note: omics_example is simulated and intended for demonstration only.

Main functions

• generate_dummy_omics() – generate linked synthetic multi-omics data
• align_omics() – align datasets by sample names
• omic_scca() – perform sparse canonical correlation analysis
• scca_to_network() – convert sCCA loadings to network edges
• plot_bipartite_network() – visualize gene–metabolite networks
• plot_pathway_circle() – feature importance radial plot
• plot_correlation_heatmap() – global correlation heatmap

Code of conduct

Please note that the OmicNetR project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.