Melody: meta-analysis of microbiome association studies for discovering generalizable microbial signatures

Abstract Standard protocols for meta-analysis of association studies are inadequate for microbiome data due to their complex compositional structure, leading to inaccurate and unstable microbial signature selection. To address this issue, we introduce Melody, a framework that generates, harmonizes,...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhoujingpeng Wei, Guanhua Chen, Zheng-Zheng Tang
Format: Article
Language:English
Published: BMC 2025-08-01
Series:Genome Biology
Subjects:
Online Access:https://doi.org/10.1186/s13059-025-03721-4
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Standard protocols for meta-analysis of association studies are inadequate for microbiome data due to their complex compositional structure, leading to inaccurate and unstable microbial signature selection. To address this issue, we introduce Melody, a framework that generates, harmonizes, and combines study-specific summary association statistics to powerfully and robustly identify microbial signatures in meta-analysis. Comprehensive and realistic simulations demonstrate that Melody substantially outperforms existing approaches in prioritizing true signatures. In the meta-analyses of five studies on colorectal cancer and eight studies on the gut metabolome, we showcase the superior stability, reliability, and predictive performance of Melody-identified signatures.
ISSN:1474-760X