TABLE 3.
Low-level: Clustering-based unsupervised integration methods.
| Approach | Clustering method | Macro category* | Author | Objective | Omics data** | Software*** |
|---|---|---|---|---|---|---|
| Kernel-based Clustering Methods | • L-MKKM (Localized Multiple Kernel K-Means) | ModE | Gönen and Margolin, (2014) | Sample-subtyping | CNV, DM, GE | • Matlab code (https://github.com/mehmetgonen/lmkkmeans) |
| • SNF (Similarity Network Fusion) | ModE | Wang et al. (2014) | Disease-subtyping | Any Omics | • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | |
| • CEPICS (https://rdrr.io/github/GaoLabXDU/CEPICS/) | ||||||
| • CancerSubtypes (https://bioconductor.org/packages/release/bioc/html/CancerSubtypes.html) | ||||||
| • rMKL-LPP (regularized Multiple Kernels Learning with Locality Preserving Projections) | ModE | Speicher and Pfeifer, (2015) | Disease-subtyping | DM, MiE, GE | • - | |
| • WSNF (Weighted SNF) | ModE | Xu et al. (2016) | Disease-subtyping | MiE, GE | • CancerSubtypes (https://bioconductor.org/packages/release/bioc/html/CancerSubtypes.html) | |
| • mixKernel | ModE | Mariette and Villa-Vialaneix, (2018) | Sample-subtyping | GE, MiE, DM | • mixKernel (https://cran.r-project.org/web/packages/mixKernel/index.html) | |
| • DSSF (Deep Subspace Similarity Fusion) | ModE | Yang et al. (2018) | Disease-subtyping | DM, MiE, GE | • - | |
| • ANF (Affinity Network Fusion) | ModE | Ma and Zhang, (2018) | Sample-subtyping | DM, MiE, GE | • ANF (https://bioconductor.org/packages/release/bioc/html/ANF.html) | |
| • NEMO (NEighborhood based Multi-Omics clustering) | ModE | Rappoport and Shamir, (2019) | Disease-subtyping | DM, MiE, GE | • NEMO (https://github.com/Shamir-Lab/NEMO) | |
| • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | ||||||
| • ab-SNF (association-signal-annotation boosted SNF) | ModE | Ruan et al. (2019) | Sample-subtyping | DM, GE | • R code (https://github.com/pfruan/abSNF/) | |
| • MvNE (Multiview Neighborhood Embedding) | ModE | Mitra et al. (2020) | Molecular-classification | DM, MiE, GE | • - | |
| • INF (Integrative Network Fusion) | DatE/ModE | Chierici et al. (2020) | Disease-subtyping, Disease-prediction | CNV, MiE, GE, PE | • Python/R code (https://gitlab.fbk.eu/MPBA/INF) | |
| • SmSPK (Smoothed Shortest Path graph Kernel) | ModE | Tepeli et al. (2020) | Sample-subtyping | GE, PE, Mutation | • Python code (https://github.com/tastanlab/pamogk) | |
| • PAMOGK (PAthway-based MultiOmic Graph Kernel clustering) | ModE | Tepeli et al. (2020) | Sample-subtyping | GE, PE, Mutation | • Python code (https://github.com/tastanlab/pamogk) | |
| (Non-negative) Matrix Factorization-based Clustering Methods | • iCluster | ModE | Shen et al. (2009) | Disease-subtyping, Biomarker-identification | CNV, GE | • iCluster (https://cran.r-project.org/web/packages/iCluster/index.html) |
| • iClusterPlus (https://bioconductor.org/packages/release/bioc/html/iClusterPlus.html) | ||||||
| • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | ||||||
| • CEPICS (https://rdrr.io/github/GaoLabXDU/CEPICS/) | ||||||
| • CancerSubtypes (https://bioconductor.org/packages/release/bioc/html/CancerSubtypes.html) | ||||||
| • jNMF (Joint Non-negative Matrix Factorization) | ModE | Zhang et al. (2012) | Disease-insight, Module-discovery | MiE, DM, GE | • - | |
| • iClusterPlus | ModE | Mo et al. (2013) | Disease-subtyping | CNV, DM, GE | • iClusterPlus (https://bioconductor.org/packages/release/bioc/html/iClusterPlus.html) | |
| Biomarker-identification | ||||||
| • FA (Factor Analysis) | DatE | Liu et al. (2013) | Disease-subtyping | MiE, GE, PE | • - | |
| • moCluster | ModE | Meng et al. (2016) | Disease-subtyping, Molecular-subtyping | MiE, DM, PE | • mogsa (https://www.bioconductor.org/packages/release/bioc/html/mogsa.html) | |
| • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | ||||||
| • JIVE (Joint and Individual Variation Explained) | ModE | O’Connell and Lock, (2016) | Disease-subtyping | MiE, DM, GE | • R.jive (https://cran.r-project.org/web/packages/r.jive/index.html) | |
| • iNMF (integrative Non-negative Matrix Factorization) | ModE | Yang and Michailidis, (2016) | Disease-subtyping | MiE, DM, GE | • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | |
| • Python code (https://github.com/yangzi4/iNMF) | ||||||
| • PFA (Pattern Fusion Analysis) | ModE | Shi et al. (2017) | Disease-subtyping | MiE, DM, GE | • - | |
| • IS -means (Integrative Sparse k-means) | DatE | Huo and Tseng, (2017) | Disease-subtyping | CNV, DM, GE | • IS-Kmeans (https://github.com/Caleb-Huo/IS-Kmeans) | |
| • MOGSA (Multi-Omics Gene-Set Analysis) | DatE | Meng et al. (2019) | Disease-insight | • GE, CNV, PE | • Mogsa (https://www.bioconductor.org/packages/release/bioc/html/mogsa.html) | |
| • SCFA (Subtyping via Consensus Factor Analysis) | ModE | Tran et al. (2020) | Disease-subtyping | • DM, MiE, GE | • R code (https://github.com/duct317/SCFA) | |
| Bayesian Clustering Methods | • TMD (Transcriptional Modules Discovery) | ModE | Savage et al. (2010) | Disease-subtyping | GE, TF | • - |
| • PARADIGM (PAthway Recognition Algorithm using Data Integration on Genomic Models) | ModE | Vaske et al. (2010) | Disease-subtyping and Disease-insight | CNV, GE, PE | • GIANT interface (http://giant.princeton.edu/) | |
| • PSDF (Patient-Specific Data Fusion) | ModE | Yuan et al. (2011) | Disease-subtyping | CNV, GE | • Matlab code (https://sites.google.com/site/patientspecificdatafusion/) | |
| • MDI (Multiple Dataset Integration) | ModE | Kirk et al. (2012) | Disease-subtyping | GE, PE | • Matlab code (https://warwick.ac.uk/fac/cross_fac/zeeman_institute/zeeman_research/software/) | |
| • BCC (Bayesian Consensus Clustering) | ModE | Lock and Dunson, (2013) | Disease-subtyping | MiE, DM, GE, PE | • bayesCC (https://github.com/ttriche/bayesCC) | |
| • LRAcluster (Low-Rank-Approximation) | ModE | Wu et al. (2015) | Disease-subtyping | CNV, DM, GE | • LRAcluster (http://lifeome.net/software/lracluster/) | |
| • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | ||||||
| Multivariate and Other Clustering Methods | • COCA (Cluster-Of-Cluster Assignment) | ModE | Hoadley et al. (2014) | Disease-subtyping | MiE, CNV, DM, GE, PE | • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) |
| • coca (https://github.com/acabassi/coca) | ||||||
| • iPF (integrative Phenotyping Framework) | DatE | Kim et al. (2015) | Sample-subtyping | MiE, GE | • iPF (http://tsenglab.biostat.pitt.edu/software.htm) | |
| • Clusternomics | ModE | Gabasova et al. (2017) | Disease-subtyping | MiE, DM, GE, PE | • Clusternomics (https://github.com/evelinag/clusternomics) | |
| • PINS (Perturbation clustering for data INtegration and disease Subtyping) | ModE | Nguyen et al. (2017) | Disease-subtyping | MiE, CNV, DM, GE | • - | |
| • iDRW (integrative Directed Random Walk) | DatE | Kim et al. (2018) | Disease-subtyping, Biomarker-discovery | DM, GE | • R code (https://github.com/sykim122/iDRW) | |
| • PINSPlus | ModE | Nguyen et al. (2019) | Disease-subtyping | MiE, CNV, DM, GE | • PINSPlus (https://cran.r-project.org/web/packages/PINSPlus/index.html) | |
| • MOVICS (https://xlucpu.github.io/MOVICS/MOVICS-VIGNETTE.html) | ||||||
| • Subtype-GAN | ModE | Yang et al. (2021) | Disease-subtyping | MiE, CNV, DM, GE | • R code (https://github.com/haiyang1986/Subtype-GAN) |
*Macro categories include (A) Multi-step and Sequential Analysis (MS-SA), (B) Data-ensemble (DatE), (C) Model-ensemble (ModE). ** CNV: copy number variation, DM: DNA methylation, MiE: Micro RNA expression, GE: gene expression, TF: transcriptional factor, PE: Protein expression. ***R packages, unless otherwise stated.