Eur. Phys. J. B (2013) 86: 250
https://doi.org/10.1140/epjb/e2013-31111-8

Regular Article

Data-driven reconstruction of directed networks

¹ Potsdam Institute for Climate Impact Research (PIK), 14412 Potsdam, Germany
² Department of Physics, Humboldt University of Berlin, 12489 Berlin, Germany
³ Interdisciplinary Center for Dynamics of Complex Systems, University of Potsdam, 14476 Potsdam, Germany
⁴ Max Planck Institute for Molecular Physiology, 44227 Dortmund, Germany
⁵ Systems Biology and Mathematical Modeling Group, Max Planck Institute for Molecular Plant Physiology, 14476 Potsdam, Germany
⁶ Institute of Biochemistry and Biology, University of Potsdam, 14476 Potsdam, Germany

^a e-mail: sabrina.hempel@pik-potsdam.de

Received: 10 December 2012
Received in final form: 20 March 2013
Published online: 6 June 2013

Abstract

We investigate the properties of a recently introduced asymmetric association measure, called inner composition alignment (IOTA), aimed at inferring regulatory links (couplings). We show that the measure can be used to determine the direction of coupling, detect superfluous links, and to account for autoregulation. In addition, the measure can be extended to infer the type of regulation (positive or negative). The capabilities of IOTA to correctly infer couplings together with their directionality are compared against Kendall’s rank correlation for time series of different lengths, particularly focussing on biological examples. We demonstrate that an extended version of the measure, bidirectional inner composition alignment (biIOTA), increases the accuracy of the network reconstruction for short time series. Finally, we discuss the applicability of the measure to infer couplings in chaotic systems.