Fits a Bayesian group-sparse multi-task regression model using Gibbs sampling. The hierarchical prior encourages shrinkage of the estimated regression coefficients at both the gene and SNP level. The model has been applied successfully to imaging phenotypes of dimension up to 100; it can be used more generally for multivariate (non-imaging) phenotypes.
|Author||Farouk S. Nathoo, Keelin Greenlaw, Mary Lesperance|
|Date of publication||2016-10-04 17:49:35|
|Maintainer||Farouk S. Nathoo <email@example.com>|