Approximating a gene regulatory network from non-sequential data

Abstract

Given non-sequential snapshots from instances of a dynamical system, we design a compressed sensing based algorithm that reconstructs the dynamical system. We formally prove that successful reconstruction is possible under the assumption that we can construct an approximate clock from a subset of the coordinates of the underlying system. As an application, we argue that our assumption is likely to be true for RNA-seq datasets, and thus we can recover the underlying nuclear receptor networks and predict pathways, as opposed to genes, that may differentiate between interesting phenotypes in some publicly available datasets.

Tasks

Reproductions