Work

scDIVA: semi-supervised integration and fine-grained annotation of tumor-immune single-cell atlases

Year: 2026

Type: software

Abstract: scDIVA adapts the Domain Invariant Variational Autoencoder to scRNA-seq for semi-supervised integration, fine-grained immune cell type annotation, and detection of out-of-reference cell states in tumor-immune single-cell atlases. Licensing. The scDIVA model and wrappers are released under the MIT license. Two paths are licensed GPL-3, inherited from upstream miloR v2.2.0: the bundled miloRVW/ fork, and milo_postprocessing_1dir.R, which contains code adapted from miloR's graphSpatialFDR. See LICENSE in the archive for the per-path map.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.21829083

Authors: Rapolu, Viraj, Karbalayghareh, Alireza, Lee, Brennan, Wong, Wilfred, Leslie, Christina

Institution: Memorial Sloan Kettering Cancer Center

Language: English

Topic: Single-cell and spatial transcriptomics

Subfield: Molecular Biology

Field: Biochemistry, Genetics and Molecular Biology

Domain: Life Sciences

Open Access status: green

OpenAlex ID: W7196939394

Cites:

Cited by:

Related to: 10