Genome-wide association studies (GWASs) have been instrumental in understanding complex phenotypic traits. However, they have rarely been used to understand lineage-specific pathways and functions that contribute to the trait. In this study, by integrating lineage-specific enhancers from mesenchymal and myeloid compartments with bone mineral density loci, we were able to segregate osteoblast- and osteoclast (OC)-specific functions. Specifically, in OCs, a PU.1-dependent transcription factor (TF) network was revealed. Deletion of PU.1 in OCs in mice resulted in severe osteopetrosis. Functional genomic analysis indicated PU.1 and MITF orchestrated a TF network essential for OC differentiation. Several of these TFs were regulated by cooperative binding of PU.1 with BRD4 to form superenhancers. Further, PU.1 is essential for conformational changes in the superenhancer region of Nfatc1. In summary, our study demonstrates that combining GWASs with genome-wide binding studies and model organisms could decipher lineage-specific pathways contributing to complex disease states.