Important biological insights can arise from the joint analysis of multiple datasets. This talk will focus on three types of joint analyses useful in genomics: signal classification, multivariate regression, and mediation analysis. Selected examples from the literature will be reviewed. New methods for each type of analysis will also be presented, building on recent developments in false discovery rate control, compound decision theory and empirical Bayes, and high-dimensional inference methodologies. These three case studies will hopefully stimulate ideas for creative analysis strategies based on combining multiple genomic datasets.