metPropagate: network-guided propagation of metabolomic information for prioritization of neurometabolic disease genes

Many inborn errors of metabolism (IEMs) are amenable to treatment, therefore early diagnosis before irreversible damage occurs is imperative. Despite recent advances, the genetic basis of many metabolic phenotypes remains unknown. For discovery purposes, Whole Exome Sequencing (WES) variant prioritization coupled with phenotype-guided clinical and bioinformatics expertise is currently the primary method used to identify novel disease-causing variants; however, it can be challenging to identify the causal candidate gene given the large number of plausible variants. Using untargeted metabolomics (UM) to prioritize metabolically relevant candidate genes is a promising approach to diagnosing known or novel IEMs in a single patient. Here, we present a network-based bioinformatics approach, metPropagate, that uses UM data from a single patient and a group of controls to prioritize candidate genes. We validate metProp on 107 patients with diagnosed IEMs and 11 patients with novel IEMs diagnosed through the TIDE gene discovery project at BC Children's Hospital. The metPropagate method ranks candidate genes by considering the network of interactions between them. This is done by using a graph smoothing algorithm called label propagation to quantify the metabolic disruption in genes' local neighbourhood. metPropagate was able to prioritize the causative gene in the top 20th percentile of candidate genes for 91% of patients with known IEM disorders. For novel IEMs, metPropagate placed the causative gene in the top 20th percentile in 9/11 patients. Using metPropagate, the causative gene was ranked higher than Exomiser's phenotype-based ranking in 6/11 patients. The results of this study indicate that for diagnostic and gene discovery purposes, network-based analysis of metabolomics data can lend support to WES gene-discovery methods by providing an additional mode of evidence to help identify causal genes.