multi-omics-integration

You can integrate multi-omics data using factor analysis, supervised classification, and network fusion methods to explore cross-modality variation and predict biological outcomes from normalized data blocks.

Identifying latent factors across modalities is best achieved through unsupervised factor analysis using MOFA+, which finds shared sources of variation driving heterogeneity in multi-modal datasets.

Yes, you need MuData containers to organize pre-normalized data blocks, ensuring consistent cross-modality integration when applying SNF for similarity-based patient stratification.

DIABLO performs supervised classification by integrating multiple omics layers to identify correlated multi-omics signatures, enabling predictive modeling of biological outcomes like disease status.

Yes, you can infer signaling topology from integration results using provided causal modeling workflows that calculate signaling networks from multi-omics variation and biological outputs.

Unsupervised factor analysis requires pre-normalized data blocks within MuData containers and may not directly predict biological outcomes without subsequent supervised classification or network fusion steps.