A1035
Title: Using taxonomic ranks improves the prediction of case-control analysis of microbiome data
Authors: Yujin Chung - Kyonggi University (Korea, South) [presenting]
Abstract: Recent studies reveal that microbial traits are differentially correlated in a phylogenetic tree. These results suggest that microbiome-based predictive models are improved by incorporating phylogenetic trees through the cophenetic distance. We propose a new way to use taxonomic ranks when a phylogenetic tree is not provided. We modified two phylogenetic tree-based predictive models to employ taxonomic ranks rather than phylogenetic trees. These predictive models were applied to microbiome data from patients with cirrhosis and hepatocellular carcinoma (HCC) and controls. The analysis shows that the taxonomic ranks improve predictive models as much as phylogenetic trees.