Locus-by-locus analysis of GWAS loci to identify the causative snp(s) and their related genes has yielded useful insights.  However, this analysis is laborious, and it does not address how the multiple loci combine in terms of function to affect susceptibility, nor how their effects interact with exposures. To address these questions, we are taking a network-based approach.  Starting with transcription factor centric gene regulatory networks, we map onto these networks the genes whose expression is altered by variation at the GWAS loci.  In breast cancer, we find that the overlapping regulons of 36 transcription factors (TFs) are enriched for genes associated with the 72 GWAS loci formally confirmed at the time of our analysis.  These regulons and their TFs show marked clustering in the network, around the TFs ESR1, FOXA1 and GATA3.  These and other data suggest that the effects at multiple GWAS loci converge on common pathways and mechanisms.  The same group of TFs are frequently involved in somatic mutation in breast cancer, suggesting that germline and somatic variation impact the same mechanisms. We are extending our analysis to lung cancer.  Here, the availability of airway epithelial cells and smoking as a defined exposure allow comparisons between individuals who do or do not smoke, and who do or do not have cancer or other respiratory disease.  Our goal is to explore the use of gene regulatory networks as a framework to integrate the effects of germline and somatic variation and exposures, and to interpret in functional terms differences in gene expression between individuals that are correlated with differences in disease phenotype.