Over the past few years, large scale genomics projects such as the ENCODE and Roadmap Epigenomics have produced genome-wide data on a large number of biochemical assays for a diverse set of human cell types and tissues. Such data play an important role in predicting the functional effects of noncoding genetic variation. We discuss here unsupervised approaches to integrate these diverse annotations for specific tissues and cell types into a single predictor of tissue-specific functional importance. We provide a global view of the sharing of functional variants across large number of tissues and cell types, and demonstrate that functional variants in promoters are more likely to be shared across many tissues compared with enhancers. A multidimensional scaling analysis based on functional scores in multiple tissues reveals clear patterns of similarity between certain tissue types, including similarity between primary tissues derived from the same embryonic tissue of origin. Using eQTL data from the Genotype-Tissue Expression (GTEx) project we show that eQTLs in specific GTEx tissues tend to be most enriched among the functional variants in relevant tissues in Roadmap. Furthermore, we show how these integrated functional scores can be used to derive the most likely tissue-/cell-type for a complex trait using summary statistics from genome-wide association studies (GWAS), and derive a tissue-based correlation map of various complex traits. Finally, we show how the tissue-specific functional scores in conjunction with GWAS summary statistics can shed light on genes and biological processes implicated in a complex trait. Functional scores are available for every possible position in the hg19 human reference genome for 127 tissues and cell types assayed in Roadmap Epigenomics.