The paper introduces MaAsLin 2, a statistical method for assessing multivariable associations between microbial community features and complex metadata in population-scale observational studies. MaAsLin 2 uses generalized linear and mixed models to handle various data types, including counts and relative abundances, with or without covariates and repeated measurements. The method was evaluated through extensive simulations under a broad range of scenarios, demonstrating its ability to preserve statistical power in the presence of repeated measures and multiple covariates while controlling false discovery rates. The software is freely available and has been applied to a microbial multi-omics dataset from the Integrative Human Microbiome Project (iHMP2) to identify biomarkers for inflammatory bowel diseases (IBD). The results show that MaAsLin 2 can detect both known and novel associations, providing a robust tool for future analysis of human-associated and environmental microbial communities.The paper introduces MaAsLin 2, a statistical method for assessing multivariable associations between microbial community features and complex metadata in population-scale observational studies. MaAsLin 2 uses generalized linear and mixed models to handle various data types, including counts and relative abundances, with or without covariates and repeated measurements. The method was evaluated through extensive simulations under a broad range of scenarios, demonstrating its ability to preserve statistical power in the presence of repeated measures and multiple covariates while controlling false discovery rates. The software is freely available and has been applied to a microbial multi-omics dataset from the Integrative Human Microbiome Project (iHMP2) to identify biomarkers for inflammatory bowel diseases (IBD). The results show that MaAsLin 2 can detect both known and novel associations, providing a robust tool for future analysis of human-associated and environmental microbial communities.