Fits successive Lasso models for several blocks of (omics) data with different priorities and takes the predicted values as an offset for the next block.
