An improved shrinkage estimator to infer regulatory networks with Gaussian graphical models

ABSTRACT

Gaussian graphical models (GGMs) are widely used to tackle the important and challenging problem of inferring genetic regulatory networks from expression data. These models have gained much attention as they encode full conditional relationships between variables, i.e. genes. As a consequence, structure learning of a GGM requires an invertible and well-conditioned covariance matrix. Unfortunately, the usual estimator---the sample covariance matrix---is ill-suited in the "small n, large p" setting characteristic of microarray data. As an alternative, [9] proposed a shrinkage estimator that is both statistically efficient and computationally fast. The effectiveness of this estimator in bioinformatics has been illustrated by [12] who successfully used it to infer genetic regulatory networks from microarray data. Unfortunately, this improved estimator requires the shrinkage intensity to be estimated from the data, which is problematic in the "small n, large p" setting. Indeed, we show that the optimal shrinkage intensity estimator used in [9, 12] is biased. We propose a parametric bootstrap approach to estimate this bias and derive a "bias-corrected" shrinkage estimator. The applicability and usefulness of our estimator are demonstrated on both simulated and real expression data.

REFERENCES

• 1: Robert Castelo , Alberto Roverato, A Robust Procedure For Gaussian Graphical Model Search From Microarray Data With p Larger Than n, The Journal of Machine Learning Research, 7, p.2621-2650, 12/1/2006
• 2: Adrian Dobra , Chris Hans , Beatrix Jones , Joseph R. Nevins , Guang Yao , Mike West, Sparse graphical models for exploring gene expression data, Journal of Multivariate Analysis, v.90 n.1, p.196-212, July 2004 [doi>10.1016/j.jmva.2004.02.009]
• 3: D. Edwards. Introduction to Graphical Modelling. Springer Texts in Statistics. Springer, second edition, 2000.
• 4: T. Hastie, R. Tibshirani, and J. Friedman. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer Series in Statistics. Springer, New York, 2001.
• 5: W. James and C. Stein. Estimation with quadratic loss. In L. M. LeCam and J. Neyman, editors, Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, pages 361--379, Berkeley, California, 1961. University of California Press.
• 6: Anton J. Kleywegt , Alexander Shapiro , Tito Homem-de-Mello, The Sample Average Approximation Method for Stochastic Discrete Optimization, SIAM Journal on Optimization, v.12 n.2, p.479-502, 2002 [doi>10.1137/S1052623499363220]
• 7: K. Kontos and G. Bontempi. Nested q-partial graphs for genetic network inference from "small n, large p" microarray data. In M. Elloumi, J. Küng, M. Linial, R. Murphy, K. Schneider, and C. Toma, editors, Proceedings of the 2nd International Conference on Bioinformatics Research and Development (BIRD 2008), number 13 in Communications in Computer and Information Science (CCIS), pages 273--287, Heidelberg, 2008. Springer.
• 8: S. L. Lauritzen. Graphical Models. Oxford Statistical Science Series. Clarendon Press, Oxford, 1996.
• 9: Olivier Ledoit , Michael Wolf, A well-conditioned estimator for large-dimensional covariance matrices, Journal of Multivariate Analysis, v.88 n.2, p.365-411, February 2004 [doi>10.1016/S0047-259X(03)00096-4]
• 10: P. Magwene and J. Kim. Estimating genomic coexpression networks using first-order conditional independence. Genome Biology, 5:R100, 2004.
• 11: Juliane Schäfer , Korbinian Strimmer, An empirical Bayes approach to inferring large-scale gene association networks, Bioinformatics, v.21 n.6, p.754-764, March 2005 [doi>10.1093/bioinformatics/bti062]
• 12: J. Schäfer and K. Strimmer. A shrinkage approach to large-scale covariance matrix estimation and implications for functional genomics. Statistical Applications in Genetics and Molecular Biology, 4(1): 32, 2005.
• 13: E. P. van Someren, L. F. A. Wessels, E. Backer, and M. J. T. Reinders. Genetic network modeling. Pharmacogenomics, 3(4): 507--525, 2002.
• 14: Adriano V. Werhli , Marco Grzegorczyk , Dirk Husmeier, Comparative evaluation of reverse engineering gene regulatory networks with relevance networks, graphical gaussian models and bayesian networks, Bioinformatics, v.22 n.20, p.2523-2531, September 2006 [doi>10.1093/bioinformatics/btl391]
• 15: J. Whittaker. Graphical Models in Applied Multivariate Statistics. John Wiley and Sons, Inc., 1990.
• 16: A. Wille and P. Bühlmann. Low-order conditional independence graphs for inferring genetic networks. Statistical Applications in Genetics and Molecular Biology, 5(1):Article 1, 2006.

INDEX TERMS

Keywords:
Gaussian graphical model, genetic regulatory network, shrinkage