In this presentation we will review different Bayesian methods for drug safety signal detection in large databases such as the FDA Adverse Events Reporting System (FAERS) database, which is a computerized information database designed to support the FDA's post-marketing safety surveillance program for drug and therapeutic biologic products. The FDA uses FAERS to monitor adverse events and medication errors that might occur with these marketed products. The Bayesian methods for signal detection include Multi-Gamma Poisson Shrinker (MGPS) and Bayesian Confidence Propagating Neural Network (BCPNN), among others. We will review these methods and present two new methods; one a simplified Bayesian (sB) method and the other that uses Dirichlet process (DP) as a prior for modeling the reporting rates of drug-adverse event combinations. The performance of the proposed methods will be evaluated using a simulation study and the methods will be applied to the FAERS database.