Inference of genetic networks from time course expression data using functional regression with lasso penalty

TL;DR

This paper introduces a method for inferring genetic regulatory networks from time course gene expression data by applying functional regression with a Lasso penalty, effectively capturing network sparsity.

Contribution

It proposes a novel approach combining functional data analysis and Lasso regularization to infer large-scale genetic networks from time series data.

Findings

Successfully inferred yeast cell-cycle network structure

Demonstrated effectiveness on real biological data

Highlighted advantages of functional regression in network inference

Abstract

Statistical inference of genetic regulatory networks is essential for understanding temporal interactions of regulatory elements inside the cells. For inferences of large networks, identification of network structure is typical achieved under the assumption of sparsity of the networks. When the number of time points in the expression experiment is not too small, we propose to infer the parameters in the ordinary differential equations using the techniques from functional data analysis (FDA) by regarding the observed time course expression data as continuous-time curves. For networks with a large number of genes, we take advantage of the sparsity of the networks by penalizing the linear coefficients with a L_1 norm. The ability of the algorithm to infer network structure is demonstrated using the cell-cycle time course data for Saccharomyces cerevisiae.