Graphical Lasso Models

Every q2-gglasso action gets one canonical demonstration on the same dataset: 50 samples × 13 ASVs from the Atacama soil study, plus five environmental covariates. The tables are small enough that you can read a whole covariance matrix at once, and each command finishes in seconds, so you can check its output by eye before trusting the same command on 300 features in the high-dimensional Atacama chapters. None of these fits is meant to be a scientific result.

Actions by chapter

Chapter	Action it owns	New parameters
Data Preparation	transform-features, calculate-covariance	transformation, pseudo_count, keep_original_id, method, bias
Single Graphical Lasso	solve-problem (sparse)	n_samples, lambda1_min/max, n_lambda1, gamma
Sparse + Low-Rank	solve-problem (latent)	latent, mu1_min/max, n_mu1
Adaptive Graphical Lasso	solve-problem (weighted)	weights, add_metadata, scale_metadata, n_cov
Regularization Paths	model selection	path_scale, lambda1_path, mu1_path, gamma
Multiple Graphical Lasso	build-groups, multi-group solve-problem	reg, lambda2_min/max, n_lambda2, non_conforming, group_array
Latent-Component PCA	pca	n_components, color_by
Summarizing a Solution	summarize	width, height, label_size
Interpretation	—	comparing the models

A machine-checkable version of this mapping lives in the Command Coverage Matrix.

Prerequisites

Populate data/ as described in Download the Tutorial Data, and confirm that qiime gglasso --help lists all six actions — see Verifying Your Installation.

Tip

Several q2-gglasso actions fail in ways that are easy to hit and hard to diagnose: pca requires a latent solution without saying so, transform-features demands a taxonomy it never reads, and --p-rank always raises. Troubleshooting collects them.