Reconstruction of metabolic networks from high-throughput metabolite profiling data: in silico analysis of red blood cell metabolism

TL;DR

This study evaluates the use of transcriptional network algorithms, specifically ARACNE, to reconstruct red blood cell metabolic networks from synthetic high-throughput metabolite data, demonstrating comparable performance to transcriptional applications.

Contribution

It introduces a novel application of transcriptional network algorithms to metabolic data and provides benchmark datasets for future research.

Findings

ARACNE performs well in metabolic network reconstruction

Synthetic data mimics real metabolic properties

Benchmark datasets are publicly available

Abstract

We investigate the ability of algorithms developed for reverse engineering of transcriptional regulatory networks to reconstruct metabolic networks from high-throughput metabolite profiling data. For this, we generate synthetic metabolic profiles for benchmarking purposes based on a well-established model for red blood cell metabolism. A variety of data sets is generated, accounting for different properties of real metabolic networks, such as experimental noise, metabolite correlations, and temporal dynamics. These data sets are made available online. We apply ARACNE, a mainstream transcriptional networks reverse engineering algorithm, to these data sets and observe performance comparable to that obtained in the transcriptional domain, for which the algorithm was originally designed.