A critical task in microbiome studies is to identify microbial features that are associated with disease outcomes. Classic statistical variable selection methods such as Lasso, SCAD and MCP are all versatile and enjoying nice asymptotic properties. Yet, they are not quite applicable to microbiome data partially due to compositionality and small sample size. Motivated by fine-mapping of the microbiome, we propose a two-step compositional knockoff filter (CKF) to provide the effective finite-sample false discovery rate (FDR) control in high-dimensional linear log-contrast regression analysis of microbiome compositional data. In the first step, we employ the compositional screening procedure to remove insignificant microbial taxa while retaining the essential sum-to-zero constraint. In the second step, we extend the knockoff filter to identify the significant microbial taxa in the sparse regression model for compositional data. We study the asymptotic properties of the proposed two-step procedure, including both sure screening and effective false discovery control. The potential usefulness of the proposed method is also illustrated with both simulation studies and real data applications.
 

compositional data,false discovery rate control,finite sample,microbial biomarker,variable selection