LinDA: linear models for differential abundance analysis of microbiome compositional data

Differential abundance analysis is at the core of statistical analysis of microbiome data. The compositional nature of microbiome sequencing data makes false positive control challenging. Here, we show that the compositional effects can be addressed by a simple, yet highly flexible and scalable, approach. The proposed method, LinDA, only requires fitting linear regression models on the centered log-ratio transformed data, and correcting the bias due to compositional effects. We show that LinDA enjoys asymptotic FDR control and can be extended to mixed-effect models for correlated microbiome data. Using simulations and real examples, we demonstrate the effectiveness of LinDA.

Automated summary

Differential abundance analysis is a crucial statistical analysis in microbiome data. The compositional nature of microbiome sequencing data poses challenges in controlling false positives. In this study, the authors propose LinDA, a simple yet flexible and scalable approach that addresses the compositional effects by fitting linear regression models on centered log-ratio transformed data and correcting the bias caused by compositional effects. They demonstrate the effectiveness of LinDA through simulations and real examples, and highlight its asymptotic FDR control and extensions to mixed-effect models for correlated microbiome data.