Despite incredible advances in the recruitment and phenotyping of patients for genomewide association studies (GWAS) in the past decade, the ability to ascertain the causes of complex disease remains a significant challenge. Recent research implies that instead of single variants of large effect, many variants of extremely small effect represent the majority of signal associated with genetic disease. To approximate these broad effects, enormous sample sizes are required. However, such recruitment is often impossible in rare diseases. To this end, we introduce a modified model of polygenic risk score (PRS) formulation incorporating protein-protein interaction network topology. This method allows the investigation of the degree to which network effects can augment existing risk scores and provides a complementing framework within which to assess the composition of existing polygenic risk measures. Secondly, we assess the ability to discern the degree to which coexisting conditions are due to similar genetic causes using the framework described above. We assess the degree to which genetic effects are detectable in a small population enriched for prematurity. We evaluate the overlap between signal for preterm birth with that of and retinopathy of prematurity (ROP), a coincident condition believed to have independent genetic causes. We believe this work is an important step toward increasing the predictive power and interpretability of genetic risk score methods, and that the evolution of such score will help inform and direct research in genetic disease to the benefit of patients and clinicians.