Research Paper Advance Articles

Figure 3. Screening for perturbations that restore a youthful transcriptional state. (A) Scatter plot of age differences for more than 39000 conditions. Positive values correspond to a pro-aging effect on the transcriptome while negative values signify a more youthful transcriptional state. Significant differences are colored yellow. Significance was determined based on an adjusted p-value of less than 0.01 (t-test, Benjamini-Hochberg correction) (B) While most of the tested conditions did not show a significant change in their predicted age, more than 1000 conditions led to a more aged transcriptional state. In contrast, a more youthful transcriptional state could be restored in 1558 conditions. Of these, the vast majority is based on chemical compounds (purple). (C) Comparison of compounds that are considered significant in DrugAge (i.e. increasing the lifespan of the treated model organism) with their counterparts in REVIVE. Compounds that showed an adjusted p-value less than 0.01 in any condition were considered significant. The comparison was restricted to compounds (i) that were included in DrugAge and (ii) for which transcriptional samples before and after treatment in any cell type were included in the assembled dataset. Each tile is annotated and colored with the number of cases falling into each category. (D) Process enrichment for 12 aging hallmark gene sets for the top 5 effectors of each process among the interventions that restore a more youthful transcriptional state. Negative values (blue) represent the rescuing of a hallmark while positive values (yellow) indicate a pro-aging effect. X- and Y-axis labels are color coded to visualize the processes that are most rejuvenated by each intervention. (E) Heatmap of predicted age differences across multiple timepoints for interventions predicted to rejuvenate cells. Higher values (yellow) represent a pro-aging effect whereas lower values (blue) indicate age reversal.