RT Journal Article
SR Electronic
T1 Personalized Annotation-based Networks (PAN) for the Prediction of Breast Cancer Relapse
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 534628
DO 10.1101/534628
A1 Thin Nguyen
A1 Samuel C. Lee
A1 Thomas P. Quinn
A1 Buu Truong
A1 Xiaomei Li
A1 Truyen Tran
A1 Svetha Venkatesh
A1 Thuc Duy Le
YR 2019
UL http://biorxiv.org/content/early/2019/01/29/534628.abstract
AB The classification of clinical samples based on gene expression data is an important part of precision medicine. However, it is not a trivial task and it is difficult to accurately predict survival outcomes and treatment responses despite advancements in the field. In this manuscript, we show how transforming gene expression data into a set of personalized (sample-specific) networks can allow us to harness existing graph-based methods to improve classifier performance. Existing approaches to personalized gene networks, based on protein-protein interactions (PPI) or population-level models, all have the limitation that they depend on other samples in the data and must get re-computed whenever a new sample is introduced. Here, we propose a novel method, called Personalized Annotation-based Networks (PAN), that avoids this limitation by using curated annotation databases to transform gene expression data into a graph. These databases organize genes into overlapping gene sets, called annotations, that we use to build a network where nodes represent functional terms and edges represent the similarity between them. Unlike competing methods, PANs are calculated for each sample independent of the population, making it a more general solution to the single-sample network problem. Using two breast cancer datasets as a case study (METABRIC and a super-set of GEO studies), we show that PAN classifiers not only predict cancer relapse better than gene features alone, but also outperform PPI and population-level graph-based classifiers. This work demonstrates the practical advantages of graph-based classification for high-dimensional genomic data, while offering a new approach to making sample-specific networks.