In order to integrate published transcriptomic data from different sources, we used an approach previously described (1) to create COLOMBOS and VESPUCCI databases. This approach allows cross-platforms integration of gene expression public data. Raw data was downloaded from GEO and Recount2. We manually curated each individual sample using the information provided in GEO and in its related paper(s).

Each expression value displayed in the data base is a contrast between the samples and the corresponding control.

Sample contrasts:
Engelen, et al. defined a condition contrast (in our case sample contrast) as the logarithmic difference between a test and a selected reference. This reference is randomly selected from the control group and used to contrast all samples from that experiment.


We used Clustergrammer developed by the Ma’ayan lab at the Icahn School of Medicine at Mount for heatmaps visualization. This tool allows interactive visualization from PumonDB data, and it is connected to Enrichr for a gen set enrichment analysis.