Fig 2. Our data-integration framework.

We used the NMTF method to integrate four different data types: germline mutations of the five subjects presented in $M^{n_s\times n_g}$, protein-protein interactions (PPIs) from BioGRID presented in $R_1^{n_g\times n_g}$, coexpressions (COEXs) from COXPRESdb presented in $R_2^{n_g\times n_g}$ and genetic interactions (GIs) from [14] presented in $R_3^{n_g\times n_g}$. M is simultaneously decomposed into the product of three lower dimensional matrix factors, P, S and G^T; molecular networks, R_i, are simultaneously decomposed into the product of three factors, G, U_i and G^T, as detailed in Materials and methods. k₁ and k₂ are the gene and the patient clusters, respectively.