Which tool do I need?

Three independent tools for single-cell multimodal data. They don't share a data format or API. Pick the one that matches your goal, then open its Methods page for the details.

Your goal	Use	What you get
Integrate multimodal data across conditions or patients, and trace results back to genes, cells, and cell types	M3	Condition-aware latent factors, patient-level predictions, gene, cell, and cell-type attributions
Compare & choose an integration method, from our benchmark or on your own data	scMultiBench	40+ methods scored with scIB metrics, rankings & bubble plots, reproduce them on your own data
Train one model for many tasks on multimodal data, classify cells, reduce dimensions, select markers, simulate cells	Matilda	Cell-type predictions, embeddings, marker genes, synthetic cells