Understanding the molecular determinants of interindividual drug response variability remains a major challenge in pharmacogenomics. Very Important Pharmacogenes (VIPs), as defined by PharmGKB, represent genes with well-established roles in drug metabolism and efficacy. However, their activity occurs within complex molecular networks that extend beyond direct pharmacogenetic associations. We constructed a VIP-centered subnetwork and applied network topology analyses, including shortest path, signal propagation, and degree centrality, to identify key nodes mediating VIP interactions. Functional enrichment, transcription factor (TF) association, and drug–gene interaction analyses were subsequently performed to characterize the biological and pharmacological context of these networks. Our results revealed a dense VIP interactome enriched in metabolic, endocrine, and signaling pathways. Notably, we identified a subset of highly connected non-VIP genes that frequently bridge canonical VIPs, termed shadow VIPs. These genes, often encoding transcriptional regulators, such as NR1I2, NR1H4, and ESR2, and more frequent in the shortest paths connecting VIPs, such as POR, APP, and GIPC1, exhibited strong associations with approved drugs, particularly hormone-related and antineoplastic agents. This suggests that shadow VIPs may act as indirect regulators of pharmacogenomic phenotypes by influencing the expression or activity of canonical VIPs. Additionally, the analysis revealed that shadow VIPs, on average, exhibit lower RVIS values than VIPs, indicating a higher intolerance to functional mutations. This suggests that shadow VIPs are under stronger selective selection, underscoring their essential biological roles. Together, these findings expand the current pharmacogenomic framework, demonstrating that drug response mechanisms emerge from a wider network of regulatory and functional interactions. Introducing the concept of shadow VIPs highlights new molecular candidates for pharmacogenetic exploration and emphasizes the value of network-based approaches in advancing precision medicine.
Melo et al. (Thu,) studied this question.