Φ-Space: continuous cell state annotation for single-cell and spatial omics studies

Installation

Installing BioConductor dependencies:

if(!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}
BiocManager::install(c("SingleCellExperiment", "scran", "scuttle", "ComplexHeatmap", "SpatialExperiment"))

Install the GitHub version of PhiSpace:

if(!requireNamespace("devtools", quietly = TRUE)) {
  install.packages("devtools")
}
devtools::install_github('jiadongm/PhiSpace/pkg')

What is Φ-Space?

Φ-Space is a computational framework for reference-based continuous annotation of cell states in single-cell and spatial multiomics data. Currently it has two modules:

Given a bulk or single-cell RNA-seq reference dataset with potentially multiple layers of phenotypes defined in the metadata (e.g. cell type and sample source), Φ-Space can phenotype on a continuum the cells and cell-like objects in the query dataset. The core of Φ-Space is continuous phenotyping based on partial least squares (PLS) regression. Compared to conventional cell type annotation methods, Φ-Space has the following strengthss

  • Identifying continuous and out-of-reference cell states;
  • Robust against batch effects;
  • Utilising bulk and multiomic refereneces and queries;
  • More suitable for exploring the spatial patterns of rare cell types.

We have applied Φ-Space to many different use cases, including

Reference Query Note Vignettes
bulk RNA-seq scRNA-seq Dendritic cell development Stemformatics DC atlas
bulk RNA-seq scRNA-seq (Perturb-seq) T cell states development Perturb-seq
scRNA-seq scRNA-seq TBC
scRNA-seq scATAC-seq requires a bimodal bridge dataset Bridge annotation
CITE-seq (scRNA+Protein-seq) CITE-seq using both modalities Covid CITE-seq
scRNA-seq supercellular spatial RNA-seq e.g. 10x Visium, Slide-seqV2 Cell type deconvolution for Visium
scRNA-seq subcellular spatial RNA-seq (imaging-based) e.g. CosMx, 10x Xenium CosMx lung cancer microenvironment
scRNA-seq subcellular spatial RNA-seq (sequencing-based) e.g. Stereo-seq, 10x Visium HD Stereo-seq AML mouse spleen with cancer lieage tracing

Cite PhiSpace

Mao, Jiadong, Choi, Jarny and Lê Cao, Kim-Anh. (2025). Φ-Space ST: a platform-agnostic method to identify cell states in spatial transcriptomics studies. bioRxiv.

Mao, Jiadong, Deng, Yidi and Lê Cao, Kim-Anh. (2024). Φ-Space: Continuous phenotyping of single-cell multi-omics data. bioRxiv.

Check out our talk for PhiSpace single-cell multiomics at mixOmics website.