API

IO: io

Tools for loading and saving Popari data and parameters.

io.save_anndata(filepath, datasets[, ...])	Save Popari state as AnnData object.
io.load_anndata(filepath)	Load AnnData object from h5ad file and reformat for Popari.

Model

Entry points for implementations of the Popari algorithm.

model.Popari(K[, replicate_names, datasets, ...])	Popari optimization model.
model.load_trained_model(dataset_path[, context])	Load trained Popari model for downstream analysis.

Components

Objects that are helpful for working with Popari.

components.PopariDataset(dataset, replicate_name)

Wrapper around AnnData object.

Analysis: tl

Functions for visualizing and evaluating Popari results.

tl.preprocess_embeddings(datasets[, ...])	Normalize embeddings per each cell.
tl.leiden(datasets[, resolution, tolerance])	Compute Leiden clustering for all datasets.
tl.compute_ari_scores(dataset, labels, ...)	Compute adjusted Rand index (ARI) score between a set of ground truth labels and an unsupervised clustering.
tl.compute_silhouette_scores(dataset, ...[, ...])	Compute silhouette score for a clustering based on Popari embeddings.
tl.compute_empirical_correlations(trained_model)	Compute the empirical spatial correlation for a feature set across all datasets.
tl.propagate_labels(trained_model, label_key)	Propagate a label from the most binned layer to the least binned layer.
tl.metagene_gsea(trained_model, metagene_index)	Run GSEA on metagenes from trained model.
tl.find_differential_genes(trained_model[, ...])	Identify genes/features that distinguish differential metagenes within a group.
tl.plot_gene_activations(trained_model, ...)	Plot metagene activation heatmaps for target genes across all groups.
tl.plot_gene_trajectories(trained_model, ...)	Plot metagene activation lineplots for target genes across all groups.
tl.evaluate_classification_task(dataset, ...)
tl.compute_confusion_matrix(dataset, labels, ...)	Compute confusion matrix for labels and predictions.
tl.compute_columnwise_autocorrelation(dataset)

Plotting: pl

Functions for visualizing and evaluating Popari results.

pl.in_situ(dataset[, axes, fig, color])	Plot a categorical label across all datasets in-situ.
pl.all_embeddings(trained_model[, ...])	Plot all learned metagenes in-situ across all replicates.
pl.metagene_embedding(datasets, metagene_index)	Plot a single metagene in-situ across all datasets.
pl.multireplicate_heatmap(datasets[, ...])	Plot 2D heatmap data across all datasets.
pl.multigroup_heatmap(trained_model[, ...])	Plot 2D heatmap data across all datasets.
pl.confusion_matrix(dataset, labels[, ax, ...])
pl.metagene_signature_enrichment(datasets, ...)	Plot enrichment of metagene signature across categories.
pl.clusters_to_categories(datasets, ...[, ...])	Plot clusters to category correspondence, based on category marker gene expression.

Simulation

Tools to generate simulated (multisample) spatially-resolved transcriptomics, or (m)SRT.

simulation_framework.SimulationParameters([...])	Container for simulation parameters.
simulation_framework.SyntheticDataset(...[, ...])	Simulated spatial transcriptomics dataset.
simulation_framework.MultiReplicateSyntheticDataset(...)	Synthetic multireplicate dataset to model biological variation and batch effects in spatial transcriptomics data.