EJP-RD WP13 case-study CAKUT momix analysis

Joint multi-omics dimensionality reduction approaches for CAKUT data using peptidome and proteome data

Brief description In (Cantini et al. 2020), Cantini et al. evaluated 9 representative joint dimensionality reduction (jDR) methods for multi-omics integration and analysis and . The methods are Regularized Generalized Canonical Correlation Analysis (RGCCA), Multiple co-inertia analysis (MCIA), Multi-Omics Factor Analysis (MOFA), Multi-Study Factor Analysis (MSFA), iCluster, Integrative NMF (intNMF), Joint and Individual Variation Explained (JIVE), tensorial Independent Component Analysis (tICA), and matrix-tri-factorization (scikit-fusion) (Tenenhaus, Tenenhaus, and Groenen 2017; Bady et al. 2004; Argelaguet et al. 2018; De Vito et al. 2019; Shen, Olshen, and Ladanyi 2009; Chalise and Fridley 2017; Lock et al. 2013; Teschendorff et al. 2018; Žitnik and Zupan 2015).

The authors provided their benchmarking procedure, multi-omics mix (momix), as Jupyter Notebook on GitHub (https://github.com/ComputationalSystemsBiology/momix-notebook) and project environment through Conda. In momix, the factorization methods are called from an R script, and parameters of the methods are also set in that script. We did not modify the parameters of the methods in the provided script. We set factor number to 2.