NEWS

gedi2 2.3.6 (2026-10-05)

Compatibility with Eigen 5.0 (upcoming RcppEigen 0.4.x). Eigen 5.0 removes Eigen::MappedSparseMatrix, so the sparse-matrix helpers compute_Yp(), compute_Mp(), compute_s_0(), compute_o_0() and eigenSparseMatVecProduct() now take Eigen::Map<Eigen::SparseMatrix<double>> instead. The package still builds against the current CRAN RcppEigen (0.3.4.0.2) and the RcppEigen release candidate (0.4.9.9-2). No user-visible changes. Contributed by Dirk Eddelbuettel (@eddelbuettel) in #28; see RcppCore/RcppEigen#151 for the wider Eigen 5.0 migration.

Count matrices are now coerced to dgCMatrix before reaching the C++ backend. Matrix >= 1.8 stores integer counts as igCMatrix, which caused CreateGEDIObject() to fail with "Need S4 class dgCMatrix for a mapped sparse matrix".

DESCRIPTION now cites the GEDI 2.0 paper (doi:10.1093/bioinformatics/btag334), and LICENSE is restored to the CRAN YEAR/COPYRIGHT HOLDER stub (full MIT text in LICENSE.md).

plot_features() gains two differential projection types (#26):
- projection = "diffexp" plots the per-cell differential expression for selected genes, given a contrast (optionally adding the global offset via include_O = TRUE). The full J x N matrix is never materialised.
- projection = "diffadb" plots the per-cell differential pathway activity for selected pathways (requires a gene-level prior C).
New method model$diffADB(contrast) returns the differential pathway activity (num_pathways x N). It is the exact differential of ADB (i.e. ADB(Z + dQ) - ADB(Z)), applying the same solve_A shrinkage so it stays on the same scale as model$projections$ADB.

Replace all cat() / print() calls with message() or warning() for suppressible console output.
Add verbose parameters to seurat_to_gedi(), gedi_to_seurat(), check_optional_dependencies(), and install_optional_dependencies().
Replace cat()-based progress bars with txtProgressBar() across imputation and training routines.
Replace installed.packages() with requireNamespace() for dependency checking.
Move hdf5r from Imports to Suggests (optional dependency for H5AD I/O).
Add ggplot2 and scales to Imports; add uwot and digest to Suggests.
Use CRAN-required two-line LICENSE format.
Add @return documentation tags to all exported and documented functions.
Replace non-ASCII characters in C++ source files with ASCII equivalents.
Remove redundant Maintainer field from DESCRIPTION.

Initial public release with C++ backend and R6 interface.
Support for multiple data modalities (count matrices, paired data, binary indicators).
Latent variable model with block coordinate descent optimization.
Dimensionality reduction, batch correction, and imputation.
Differential expression and pathway association analysis.
H5AD file I/O for Python interoperability.
Seurat and SingleCellExperiment integration.