TITLE: Covariance Selection and Gene Networks

CONTACT: Alexandre d'Aspremont, aspremon@ens.fr

ABSTRACT: Covariance Selection seeks to infer networks of dependence in multivariate data sets by solving a penalized maximum likelihood problem. It is of particular interest in genomics where it can be used to identify gene regulation networks from DNA sequencing data sets over large populations. 

Covariance selection algorithms have made a lot of progress compared to early methods developed about a decade ago. Furthermore, the cost of sequencing has also collapsed so significantly larger data sets are now starting to become available. This presents a unique opportunity to start testing both the numerical performance and the biological significance of these methods on realistic data.

REFERENCES:

[1] Hsieh, C. J., Sustik, M. A., Dhillon, I. S., & Ravikumar, P. (2014). QUIC: quadratic approximation for sparse inverse covariance estimation. The Journal of Machine Learning Research, 15(1), 2911-2947.

[2] d'Aspremont, A., Banerjee, O., & El Ghaoui, L. (2008). First-order methods for sparse covariance selection. SIAM Journal on Matrix Analysis and Applications, 30(1), 56-66.

[3] Schäfer, J., & Strimmer, K. (2005). A shrinkage approach to large-scale covariance matrix estimation and implications for functional genomics. Statistical applications in genetics and molecular biology, 4(1).