PatientProfiler: A network-based approach to personalized medicine

Abstract

Deciphering the intricate mechanisms underlying reprogramming in cancer cells is a crucial challenge in oncology as it holds the key to advance our ability to diagnose and treat cancer patients. For this reason, comprehensive and patient-specific multi-omic characterization of tumor specimens has become increasingly common in clinical practice. While these efforts have advanced our understanding of the molecular mechanisms underlying breast cancer progression, the identification of personalized therapeutic approaches remains a distant goal. The main shortcoming is the absence of a robust computational framework to integrate and interpret the available multi-dimensional data and to drive translational solutions. To fill this gap, we developed PatientProfiler, a computational pipeline that leverages causal interaction data, annotated in our in-house manually-curated resource, SIGNOR, to address how the genetic and molecular background of single patients contributes to the establishment of a malignant phenotype. PatientProfiler is an open-source, R-based package composed of several functions that allows multi-omic data analysis and standardization, generation of patient-specific mechanistic models of signal transduction, and extraction of network-based prognostic biomarkers. To benchmark the tool, we retrieved genomic, transcriptomic, (phospho)proteomic, and clinical data derived from 122 treatment-naive breast cancer biopsies, available at the CPTAC portal. Thanks to this approach, we identified patient-specific mechanistic models (one patient, one network) that recapitulate dysregulated signaling pathways in breast cancer. This collection of models provides valuable insights into the underlying mechanisms of tumorigenesis and disease progression. Moreover, in-depth topological exploration of these networks has allowed us to define seven communities (subnetworks), each associated with a unique transcriptomic signature and a distinct prognostic value. In summary, our work demonstrates that PatientProfiler is a tool for patient-specific network analysis, advancing personalized medicine towards the identification of actionable biomarkers and tailored therapeutic strategies.