Lab 07 – Clustering and Classification of Microbiome Data

0.1 Motivation

There are a growing number of microbiome applications for which it appears the microbiome is predictive of some host/patient/environment outcome, or improves the predictability of some outcome when used in conjunction with additional sample covariate(s).

0.2 Goals of this Lab

The goal of this lab is to show some examples

• 1. Unsupervised learning, mainly evaluating the number of clusters in an example dataset, and
• 2. Supervised learning, training a predictive model based on SVM or Random Forest, and estimate its predictive performance using cross-validation.