Fig. 2.

Identification of clinically viable compounds. a Schematic workflow of the in silico search for clinically viable drugs based on the identified DE genes. Briefly, the colon to metastasis disease gene signature (Table 1) was used to query the CMAP database, a collection of paired gene expression profiles from ctrl- and drug-treated cell lines. A positive correlation denotes the degree of similarity and a negative correlation emphasizes an inverse similarity between the query signature and the reference profile generated by the chemical perturbation. b Pie chart showing the distribution of the most represented class of compounds identified among those more negatively correlated with our bait signature (connectivity score between -100 and -50). c Representative histogram showing all the identified compounds ranked by the CMAP connectivity score. On y-axis is reported the connectivity score with negative values indicating negative-correlation and positive value indicating positive correlation. Lines indicate the approximate position of the four compounds selected for further testing