## Softwares## EiGLasso[Download]EiGLasso: Scalable Estimation of Cartesian Product of Sparse Inverse Covariance Matrices. J. Yoon and S. Kim. Uncertainty in Artificial Intelligence (UAI), 2020. ## PerturbNet[Download]This software implments the statistical methods for learning gene networks underlying clinical phenotypes using SNP perturbations, described in the following paper: Learning gene networks underlying clinical phenotypes using SNP perturbation. C. McCarter, J. Howrylak, S. Kim. PLoS computational biology, 16(10), e1007940, 2020. The statistical model and optimization method were previously described in the following machine learning conference papers: Large-scale optimization algorithms for sparse conditional Gaussian graphical models. C. McCarter, S. Kim. In Proceedings of the Conference on Artificial Intelligence and Statistics (pp. 528-537), PMLR, 2016. On sparse Gaussian chain graph models. C. McCarter, S. Kim. Advances in Neural Information Processing Systems, 27, 3212-3220, 2014. ## Sparse Conditional Gaussian Graphical Models[Download]PerturbNet above provides two to three orders of magnitude faster optimization. Sparse conditional Gaussian graphical model (CGGMs) and its learning algorithm decode the SNP perturbation effects of gene regulatory system captured in population gene expression and genotype data to learn a gene network along with expression quantitative trait loci (eQTLs) that perturb this network. The software package includes an implementation of the learning/inference algorithms for sparse CGGMs described in the following papers: Learning gene networks under SNP perturbations using eQTL datasets. L. Zhang, S. Kim. PLoS Computational Biology, 6(3):1095-1117, 2014. Joint estimation of structured sparsity and output structure in multiple-output regression via inverse-covariance regularization. K.A. Sohn, S. Kim. Proceedings of the 15th International Conference on Artificial Intelligence and Statistics (AISTATS), 2012. ## A* Lasso[Download]A* lasso for learning a sparse Bayesian network structure. J. Xiang, S. Kim. Advances in Neural Information Processing Systems (NIPS), 2013 ## Tree-Guided Group Lasso[Download]Tree-guided group lasso estimates a sparse multi-response regression model, while leveraging a hierarhical clustering tree structure over response variables. In particular, given a hierarchical structure over the responses such as a hierarchical clustering tree with leaf nodes for responses and internal nodes for clusters of related responses at multiple granularity, tree-guided group lasso leverages this structure to recover covariates relevant to each hierarchically-defined cluster of responses. The method can be used for expression quantitative trait locus (eQTL) mapping to identify genetic variants perturbing the expressions of gene modules defined by hierarchical clustering tree. The software package includes an implementation of smoothing proximal gradient method described in the papers below: Tree-guided group lasso for multi-response regression with structured sparsity, with an application to eQTL mapping. S. Kim, E. P. Xing. Annals of Applied Statistics, 6(3):1095-1117, 2012. Smoothing proximal gradient method for general structured sparse regression. X. Chen, Q. Lin, S. Kim, J.G. Carbonell, E.P. Xing. Annals of Applied Statistics, 6(2):719-752, 2012. ## Graph-Guided Fused Lasso[Download]Graph-guided fused lasso estimates a sparse multi-response regression model, while leveraging a weighted network structure over response variables to find covariates that are jointly relevant to multiple correlated responses according to the network structure. The method can be used for expression quantitative trait locus (eQTL) mapping or genome-wide association analysis, where the network structure over phenotypes is known. The software package includes an implementation of smoothing proximal gradient method described in the papers below: Smoothing proximal gradient method for general structured sparse regression. X. Chen, Q. Lin, S. Kim, J.G. Carbonell, E.P. Xing. Annals of Applied Statistics, 6(2):719-752, 2012. [Home] |