RT Journal Article
SR Electronic
T1 Integrative pharmacogenomics to infer large-scale drug taxonomy
JF bioRxiv
FD Cold Spring Harbor Laboratory
SP 046219
DO 10.1101/046219
A1 Nehme El-Hachem
A1 Deena M.A. Gendoo
A1 Laleh Soltan Ghoraie
A1 Zhaleh Safikhani
A1 Petr Smirnov
A1 Ruth Isserlin
A1 Gary D. Bader
A1 Anna Goldenberg
A1 Benjamin Haibe-Kains
YR 2016
UL http://biorxiv.org/content/early/2016/04/06/046219.abstract
AB Identification of drug targets and mechanism of action (MoA), particularly for new and uncharacterized drugs, is important for the optimization of drug efficacy. Current approaches towards determining drug MoA largely rely on prior information such as side effects, therapeutic indication and chemo-informatics. However, such information is not transferable or applicable for newly identified small molecules. Despite continuous release of large-scale pharmacogenomic datasets, these valuable data remain underused to classify drugs. Accordingly, a systematic and unbiased approach towards MoA prediction is imperative to efficiently classify new compounds and infer their potential targets of MoA. Here, we propose a method that only relies on basic drug characteristics, including drug structural information, drug perturbation and drug sensitivity profiles, which have not been previously combined towards predicting drug targets and MoA. We harnessed the full potential of pharmacogenomics data using our Similarity Network Fusion approach to implement Drug Network Fusion (DNF), a scalable, integrative drug taxonomy. We demonstrate that DNF is effective towards prediction of drug targets and anatomical therapeutic chemical (ATC) classification). Our method enables robust inference of drug MoAs for new and existing compounds, using integrative computational pharmacogenomics