Pathway-based subnetworks enable cross-disease biomarker discovery

Syed Haider ,Píetro Lió ,Maud H.w Starmans ,Luc Dirix ,Annette Hasenburg ,Christos Markopoulos ,Cassandra Brookes ,Cheryl Crozier ,Xihui Lin ,Nathalie C Moon ,Cornelis J.h Van De Velde ,Victoria Sabine ,Camilla Drake ,Arek Kasprzyk ,Vincent Stimper ,Jianxin Wang ,Cindy Q Yao ,John M.s Bartlett ,Daniel Rea ,Dirk G Kieback ,Philippe Lambin ,Michal R Grzadkowski ,Francis Nguyen ,Caroline Seynaeve ,Paul C Boutros

doi:10.1038/s41467-018-07021-3

Abstract

Biomarkers lie at the heart of precision medicine. Surprisingly, while rapid genomic profiling is becoming ubiquitous, the development of biomarkers usually involves the application of bespoke techniques that cannot be directly applied to other datasets. There is an urgent need for a systematic methodology to create biologically-interpretable molecular models that robustly predict key phenotypes. Here we present SIMMS (Subnetwork Integration for Multi-Modal Signatures): an algorithm that fragments pathways into functional modules and uses these to predict phenotypes. We apply SIMMS to multiple data types across five diseases, and in each it reproducibly identifies known and novel subtypes, and makes superior predictions to the best bespoke approaches. To demonstrate its ability on a new dataset, we profile 33 genes/nodes of the PI3K pathway in 1734 FFPE breast tumors and create a four-subnetwork prediction model. This model out-performs a clinically-validated molecular test in an independent cohort of 1742 patients. SIMMS is generic and enables systematic data integration for robust biomarker discovery.

Highlights

Biomarkers lie at the heart of precision medicine
While the N model assumes independent and additive effects of parts of a subnetwork, the E and N + E models incorporate the impact of dysregulated interactions (Methods)
SIMMS fits each one of these models thereby estimating a ‘module-dysregulation score’ (MDS) for each subnetwork that measures their strength of association with a specific disease, phenotype or outcome (Supplementary Fig. 1)

Summary

Introduction

Biomarkers lie at the heart of precision medicine. Surprisingly, while rapid genomic profiling is becoming ubiquitous, the development of biomarkers usually involves the application of bespoke techniques that cannot be directly applied to other datasets. To demonstrate its ability on a new dataset, we profile 33 genes/nodes of the PI3K pathway in 1734 FFPE breast tumors and create a foursubnetwork prediction model This model out-performs a clinically-validated molecular test in an independent cohort of 1742 patients. Several groups have integrated multiple data types using network and systems biology approaches identifying patient subtypes, with limited post-hoc clinical evaluation[7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25] These algorithms have not yet clearly shown how the interplay between different pathways underpins disease etiology, nor generated biomarkers with systematically demonstrated reproducibility on independent patient cohorts across multiple indications[26]. We validate SIMMS across five tumor types and 11,392 patients, using it to create biomarkers from a diverse range of molecular assays and uncovering unanticipated pan-cancer similarities

Methods

Results

Conclusion