scdecon¶
Single-cell-reference deconvolution of bulk tumour transcriptomes.
Bulk RNA-seq measures the average expression of a whole tissue — a tumour biopsy is a blend of cancer cells, T cells, B cells, macrophages, fibroblasts, and more. scdecon estimates what fraction of a bulk sample comes from each cell type, using an annotated single-cell RNA-seq atlas as the reference.
It builds a cell-type signature matrix from single-cell data, then solves a
constrained regression to infer cell-type proportions in bulk samples —
Bulk ≈ Signature × Proportions, subject to non-negativity (p ≥ 0) and
sum-to-one (Σ p = 1).
What it does¶
- Reference → signature. Load an annotated
.h5ad, run QC + normalisation, select marker genes, and build a linear-scale signature matrix. - Deconvolve bulk. Estimate per-sample cell-type proportions with a choice of
solvers (NNLS, ν-SVR, robust) behind one
Solverinterface. - Validate against ground truth it generates itself. Simulate pseudobulk with known proportions, then measure recovery (RMSE / Pearson / Spearman).
- Benchmark solvers fairly on one shared pseudobulk set.
- Apply to real data. A reproducible melanoma (TCGA-SKCM + Tirosh) workflow
lives in
scripts/(see the tutorial).
Where to go next¶
- Getting started — install, then run the pipeline from the command line or via Snakemake.
- CLI reference — every command, its configuration, and the exit-code policy.
- API reference — the Python package, module by module.
Scope
scdecon is research software. Quantitative accuracy is validated on self-generated synthetic ground truth; real-data results (e.g. the melanoma workflow) are relative cross-platform estimates, not absolute calibrated fractions. See the tutorial for the caveats.