Yihui Luan

This paper evaluates 22 normalization methods for predicting quantitative phenotypes in metagenomic data across 3 simulations and 31 real datasets, finding no significant superiority among them.

This study compares normalization methods for metagenomic data to improve phenotype prediction across studies, finding that batch correction methods perform best while population and disease effects limit normalization effectiveness.