Table 2.
The widely used subset of methods for three types of multi-omics integration
| Integration type | Method | Principle | URL | Language | Release year |
|---|---|---|---|---|---|
| Horizontal integration | Harmony | Iterative clustering in dimensionally reduced space | https://github.com/welch-lab/liger | R | 2019 |
| scVI | Bayesian variational autoencoder with a probabilistic formulation | https://scvi-tools.org | Python | 2021 | |
| LIGER with online iNMF | Integrative non-negative matrix factorization and joint clustering + online learning | https://github.com/welch-lab/liger | R | 2021 | |
| Vertical integration | Seurat v4 | Weighted nearest neighbor analysis | https://github.com/satijalab/seurat | R | 2021 |
| MOFA+ | Generalization of canonical correlation analysis that builds on the Bayesian group factor analysis framework | https://github.com/bioFAM/MOFA2 | Python; R | 2020 | |
| Diagonal integration | MATCHER | Gaussian process latent variable model | https://github.com/jw156605/MATCHER | Python | 2017 |
| MMD-MA | Embedding measured in different ways into a learned latent space | https://bitbucket.org/noblelab/2019_mmd_wabi/src/master/ | Python | 2019 |