Package index • quasar

QUASAR

Uncertainty-aware cell-type deconvolution. Returns Dirichlet-based point estimates together with confidence and prediction intervals.

quasar_sim_bulk(): Simulate pseudo-bulk or spatial transcriptomics profiles from single-cell data

Scaden

Native R implementation of Scaden (Menden et al., 2020). A three-model multilayer perceptron ensemble trained on simulated pseudo-bulks.

scaden_sim_pb(): Simulate pseudobulk training data for Scaden
scaden_process(): Process simulated and bulk data for Scaden
scaden(): Train a Scaden ensemble in torch
scaden_predict(): Predict cell-type proportions with a trained Scaden ensemble

TAPE

Native R implementation of TAPE (Chen et al., 2022). An autoencoder with a tissue-adaptive refinement stage, in overall or high-resolution mode.

tape_simulate(): Simulate TAPE training pseudobulks from single-cell data
tape_process(): Process simulated and bulk data for TAPE
tape_train(): Train the TAPE autoencoder
tape_predict(): Predict cell fractions and signature matrices with TAPE
tape(): Run the full TAPE workflow
tape_clone_model(): Clone a trained TAPE model

DISSECT

Native R implementation of DISSECT (Khatri et al., 2024). Semi-supervised consistency regularization for cell-type fractions and cell-type-specific expression.

dissect_simulate(): Simulate DISSECT training mixtures from single-cell data
dissect_process(): Prepare DISSECT input matrices
dissect_prop(): Estimate cell-type proportions with DISSECT
dissect_expr(): Estimate cell-type-specific expression with DISSECT
dissect(): Run the full DISSECT workflow

OmicsTweezer

Native R implementation of OmicsTweezer (Yang et al., 2025). Domain adaptation between simulated pseudo-bulks and real bulk samples.

omics_simulate(): Simulate pseudobulk training data for OmicsTweezer
omics_process(): Process simulated and bulk data for OmicsTweezer
omics_create_model(): Create an OmicsTweezer model
omics_train(): Train one OmicsTweezer model
omics_predict(): Predict cell-type proportions with one OmicsTweezer model
omics_tweezer(): Run the complete OmicsTweezer workflow
omics_set_seed(): Set reproducible random seeds for OmicsTweezer

Evaluation

Metrics for comparing estimated proportions against known ground truth.

quasar_prop_metrics(): Calculate proportion-estimation performance metrics

Normalisation

Gene filtering and sample-wise normalisation shared across methods.

normalize_bulks(): Normalize reference and target bulk matrices

Interoperability

Reading and writing AnnData .h5ad files.

quasar_h5adimport(): Import an .h5ad file as a Seurat or SingleCellExperiment object
quasar_h5adexporter(): Export a Seurat, SingleCellExperiment or bulk matrix to .h5ad
quasar_h5_read_dataframe(): Read an anndata dataframe group (obs / var) into a data.frame
quasar_h5_read_matrix(): Read an anndata matrix group / dataset (X, raw/X, layers, obsm) as an R matrix