q2-classo

A QIIME 2 plugin for solving constrained sparse regression and classification problems with microbiome data including:

• Sparse log-contrast regression
• Cross-validation for hyperparameter selection
• Stability selection for feature selection
• Classification and regression tasks
• Tree-aggregated predictive modeling (trac)
• Interactive visualizations with model diagnostics

📂 Tutorial & Examples

---

Installation

First, make sure you have the required dependencies installed:

conda install -c conda-forge zarr plotly

OR

pip install zarr plotly c-lasso

Next to install the plugin:

pip install git+https://github.com/bio-datascience/q2-classo-latest.git
qiime dev refresh-cache

---

Usage Tutorial

A complete tutorial on using q2-classo for microbiome data analysis — including preprocessing, CLR transformation, constrained regression, and visualization — is available in the repository examples.

👉 Quick Start Guide

This tutorial includes:

• CLR transformation and taxonomic aggregation
• Constrained lasso regression with cross-validation
• Stability selection for robust feature selection
• Interactive visualization of model results

---

Citation

If you use q2-classo, please cite:

Bien, J., Yan, X., Simpson, L. and Müller, C. L. (2020). Tree-Aggregated Predictive Modeling of Microbiome Data. arXiv preprint arXiv:2002.08698.

Related Projects

• c-lasso: Python solvers for constrained lasso problems
• q2-gglasso: QIIME 2 plugin for graphical lasso problems
• QIIME 2: Extensible microbiome analysis platform

---

License

BSD 3-Clause License. See LICENSE for details.