Fig. 3: PC-corr method to unveil how PPI is affecting the microbiota in gastric environment in dyspeptic patients.

(Middle panel) To investigate the effect of PPIs on the gastric microbiota in dyspeptic patients, we constructed the conserved PC-corr network at 0.5 cut-off, by merging the PC-corr networks obtained from the gastric mucosa (Paroni Sterbini et al.²²) and the gastric fluid (Amir et al.²¹). To do so, we firstly considered the union of the two PC-corr networks obtained from the gastric tissue dataset and then we intersected it with the PC-corr network from the gastric fluid dataset. All the bacteria spotted in the conserved PC-corr network (violet circle) were found increased with PPI use. In both the two studied datasets, red nodes indicate bacteria whose abundance is increased with PPI treatment, while black nodes indicate bacteria with lower abundance following treatment with this acid suppressing medication. The common bacteria that showed an opposite trend in the two datasets, i.e. microbial abundance increased in one dataset and decreased in the other dataset, were removed from the network. (Top panel) The top panel shows the obtained Amir4’s network, not in common with the Paroni Sterbini’s network. The module on the left side (except Enterobacteriaceae) include bacteria more abundant following PPI treatment in Amir4’s data, while the module on the right (and Enterobacteriacea) is composed of decreased bacteria in abundance under PPI therapy in Amir4’s data. (Bottom panel) The bottom panel represents the part of Paroni Sterbini’s network (union of the two PC-corr network), that is not shared with Amir4’s one. As in the top and middle panels, the colour of the nodes represents if the bacteria display higher (red nodes) or lower abundance (black nodes) in PPI-treated samples of Paroni Sterbini’s dataset.