Fig. 7: Dynamic network guides selection of therapeutic intervention in mouse model.
From: Crosstalk with lung fibroblasts shapes the growth and therapeutic response of mesothelioma cells
A Diagram representing the generation of diffusion profiles for drugs and phenotypes using a random walk-based network propagation algorithm. Upon convergence, the diffusion profile measures how often each node in the network is visited, revealing the most affected proteins. B Results of Spearman correlation between drugs and phenotypes diffusion profiles. Green colour highlights higher, and red colour lower, correlation. Sara saracatinib, Vistu vistusertib, Cedi cediranib, Cis cisplatin, Pem pemetrexed. C Table of p-values of correlations in (B). Each cell contains the p-value that correlations between a drug in row and phenotypes are lower than correlations between a drug in column and phenotypes. P-values confirm that cediranib, cisplatin, and pemetrexed have significantly lower correlation with phenotypes than saracatininb and vistusertib. D 20 first Reactome terms from GSEA analysis associated with the network propagation for Vistusertib, Saracatinib and the proteomic data for MMP or AF at day 3 of CM treatment. Numbers represent the ranking of the process for each category. E Saracatinib treatment increases the survival of asbestos-exposed Nf2/Bap1/Cdkn2a triple-floxed mice over cisplatin/pemetrexed (Cis + Pem) combination. Statistics: Log-rank Mantel–Cox test between Vehicle and Cis + Pem.
