Abstract: The present disclosure describes systems and methods for multivarlate analysis of adverse event data. According to a first aspect, patient-specific genomic Information is used to optimize or de-risk therapy for the patient. According to other aspects of the invention, unknown drug targets are identified via adverse event data. According to still other espects, a medication is identified to exclude from use for an indication or from a clinical trial of another medication. According to another aspect, a predicted side effect profile is generated for a medication targeting a novel target. According to still another aspect, combination therapies are identified via adverse event data. According to another aspect, molecular interactions between a plurality of molecular entities are displayed in an intuitive format. According to still another aspect, molecular entities responsible for adverse event differences between similar indications are identified.