Casual Compressive Sensing for Gene Network Inference

Abstract

We propose a novel framework for studying causal inference of gene interactions using a combination of compressive sensing and Granger causality techniques. The gist of the approach is to discover sparse linear dependencies between time series of gene expressions via a Granger-type elimination method. The method is tested on the Gardner dataset for the SOS network in E. coli, for which both known and unknown causal relationships are discovered.

References in corpus (3)

• ARACNE: An Algorithm for the Reconstruction of Gene Regulatory Networks in a Mammalian Cellular Context
• NESTA: A Fast and Accurate First-order Method for Sparse Recovery
• A Computational Algebra Approach to the Reverse Engineering of Gene Regulatory Networks

Cited by in corpus (2)

• Large-scale nonlinear Granger causality: A data-driven, multivariate approach to recovering directed networks from short time-series data
• Granger Causality for Compressively Sensed Sparse Signals