Gene Regulatory Networks from Multifactorial Perturbations Using Graphical Lasso: Application to the DREAM4 Challenge

• Published in: PloS one Vol. 5; no. 12; p. e14147
• Main Authors: Menéndez, P, Kourmpetis, I.A, Braak, C.J.F. ter, Eeuwijk, F.A. van
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 20.12.2010; Public Library of Science (PLoS)
• Subjects: Algorithms; Artificial intelligence; Bayes Theorem; Biology; Computational Biology; Computational Biology - methods; Computational Biology/Systems Biology; Computational Biology/Transcriptional Regulation; Computer applications; Covariance; Data processing; Gene expression; Gene Expression Profiling; Gene Regulatory Networks; Genes; inference; Information systems; International conferences; likelihood; Markov analysis; Markov Chains; Markov processes; Mathematics/Algorithms; Mathematics/Statistics; Models, Statistical; Multivariate Analysis; Natural language processing; Network topologies; Networks; Normal Distribution; Random variables; regression; Reproducibility of Results; selection; Topology; Training; New York; Netherlands
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0014147

A major challenge in the field of systems biology consists of predicting gene regulatory networks based on different training data. Within the DREAM4 initiative, we took part in the multifactorial sub-challenge that aimed to predict gene regulatory networks of size 100 from training data consisting of steady-state levels obtained after applying multifactorial perturbations to the original in silico network. Due to the static character of the challenge data, we tackled the problem via a sparse Gaussian Markov Random Field, which relates network topology with the covariance inverse generated by the gene measurements. As for the computations, we used the Graphical Lasso algorithm which provided a large range of candidate network topologies. The main task was to select the optimal network topology and for that, different model selection criteria were explored. The selected networks were compared with the golden standards and the results ranked using the scoring metrics applied in the challenge, giving a better insight in our submission and the way to improve it.Our approach provides an easy statistical and computational framework to infer gene regulatory networks that is suitable for large networks, even if the number of the observations (perturbations) is greater than the number of variables (genes).