Multimodal overview#

pyXenium.multimodal is the canonical public surface for joint RNA + protein work.

It is organized into four layers:

  • loading: prepare an AnnData with aligned RNA, protein, spatial coordinates, and optional cluster attachments.

  • analysis: core joint analysis tools such as rna_protein_cluster_analysis, protein_gene_correlation, and ProteinMicroEnv.

  • immune_resistance: marker/state/pathway discordance and spatial immune-resistance scoring.

  • workflows: packaged renal validation and pilot workflows with markdown, JSON, CSV, and figure artifacts.

Canonical entrypoints#

  • load_rna_protein_anndata

  • rna_protein_cluster_analysis

  • protein_gene_correlation

  • ProteinMicroEnv

  • annotate_joint_cell_states

  • compute_rna_protein_discordance

  • build_spatial_niches

  • score_immune_resistance_program

  • aggregate_multi_sample_study

  • run_validated_renal_ffpe_smoke

  • run_renal_immune_resistance_pilot

Compatibility note#

Older multimodal import paths under pyXenium.analysis, pyXenium.validation, and pyXenium.io.load_xenium_gene_protein(...) remain available for compatibility only. New code should import from pyXenium.multimodal.