Fig. 6: Personalized comorbidity profiles.

Two steps are necessary to generate patient-specific comorbidity profiles with associated drugs. a Generate a drug profile for each disease. Based on patient-specific differential expression profiles, associate to each patient the drugs that generate expression changes similar or opposite to the ones observed in the patient based on LINCS. Generate a drug profile calculating for each drug within each disease the percentage of patients to whom they are positively associated. Do the same with all the diseases. b Generate a personalized comorbidity profile. For each patient, based on molecular similarities, calculate the relative risks of developing secondary diseases based on relative molecular similarities. Associate the patient with the drugs that generate expression changes similar and opposite to the ones observed in the patient using LINCS. Select the overlap between drugs associated with the patient and those associated with each disease. Of those, filter the first-line treatments. The resulting drugs are the ones that might be involved in the increased risk of developing the detected secondary diseases.