SUMO 0.2.0
🔧 Major Enhancements
- Introduced
demo_multiomics_analysis() for end-to-end
demonstration of MOFA-based integrative analysis using both
SUMO-generated and real-world datasets (e.g., CLL).
- Added PowerPoint reporting with automatic visualization export using
officer and rvg.
🆕 New Functions
compute_means_vars(): Computes sample- and
feature-level statistics for benchmarking and noise calibration across
multiple omics datasets.plot_weights(): Refactored with modular plotting
helpers and new options for:
show.legend (toggle legend)"integrated" mode (combined view of both omics)- Signal region annotation from
sim_object$signal_annotation
plot_factor(): Now supports both scatter and histogram
visualization of factor scores. Includes support for plotting all
factors and controlling legend display.
🎨 Visualization Improvements
- Modularized plotting into helper functions
(
build_scatter_plot(),
build_histogram_plot()).
- Added consistent theming, Viridis color scales, and better axis
labeling across plots.
- Improved annotation of signal-vs-noise regions for clarity in
benchmarking.
🧪 Simulation Engine Upgrades
- Updated
OmixCraftHD() to support:
- Per-factor specification of means and variances for samples and
features (e.g.,
signal.samples = c(3, 0.5))
- Automatic derivation of signal masks used for evaluation and
annotation
- Better error handling and defaults for edge cases (e.g., NULL
factors)
🧼 Internal Cleanup
- Removed deprecated argument
signal_vector
- Added
globalVariables() suppressions for clean CRAN
checks
- Improved code readability and consistency across plotting
functions
📦 Infrastructure
- Package now includes robust examples for every major exported
function
- Improved documentation headers for
Roxygen2 and pkgdown
compatibility