A Bayesian regression and multiple changepoint model for systems biology

Abstract

We propose a Bayesian regression and multiple changepoint model for reverse engineering gene regulatory networks from high-throughput gene ex- pression profiles. We report results from a recently held international gene net- work reconstruction competition, in which our method was objectively assessed in a blind study. While we did not win the competition, the scores indicate that the proposed method favourably compares with the majority of competing ap- proaches and clearly belongs to the group of highest-ranked performers.