A0221
Title: Multigroup analysis of compositions of microbiomes with covariate adjustments and repeated measures
Authors: Huang Lin - University of Maryland (United States) [presenting]
Abstract: Microbiome differential abundance analysis for two-group comparisons is extensively documented in existing literature. However, many microbiome studies encompass more than two groups, often including ordered categories like disease stages, necessitating varied comparison approaches. Traditional pairwise comparisons fall short in terms of statistical power and control of false discovery rates. Furthermore, many studies involve repeated measures from the same participants, as seen in longitudinal microbiome research, yet there is a notable gap in the literature for effectively addressing these complex scenarios. To bridge this gap, ANCOM-BC2 is introduced, a general framework designed for multigroup analyses, accommodating covariate adjustments and repeated measures. ANCOM-BC2 has proven its efficacy in enhancing power and reducing false discovery rates when compared to competing methods in simulation studies. The methodology is also exemplified through two real-world datasets: one investigating the impact of aridity on soil microbiomes and the other examining the microbiome alterations in patients with inflammatory bowel disease post-surgical interventions.
