Inferring Time-Delayed Gene Regulatory Networks Using Cross-Correlation and Sparse Regression

摘要

Inferring a time-delayed gene regulatory network from mi-croarray gene-expression is challenging due to the small numbers of time samples and requirements to estimate a large number of parameters. In this paper, we present a two-step approach to tackle this challenge: first, an unbiased cross-correlation is used to determine the probable list of time-delays and then, a penalized regression technique such as the LASSO is used to infer the time-delayed network. This approach is tested on several synthetic and one real dataset. The results indicate the efficacy of the approach with promising future directions.