Abstract
High-throughput sequencing technologies have revolutionized microbiome research by allowing the relative quantification of microbiome composition and function in different environments. One of the main goals in microbiome analysis is the identification of microbial species that are differentially abundant among groups of samples, or whose abundance is associated with a variable of interest. Most available methods for microbiome abundance testing perform univariate tests for each microbial species or taxa separately, ignoring the compositional nature of microbiome data.
We propose an alternative approach for microbiome abundance testing that consists on the identification of two groups of taxa whose relative abundance, or balance, is associated with the response variable of interest. This approach is appealing, since it has direct translation to the biological concept of ecological balance between species in an ecosystem. In this work, we present selbal, a greedy stepwise algorithm for balance selection. We illustrate the algorithm with 16s abundance data from an HIV-microbiome study and a Crohn-microbiome study. Importance: A more meaningful approach for microbiome abundance testing is presented. Instead of testing each taxon separately we propose to explore abundance balances among groups of taxa. This approach acknowledges the compositional nature of microbiome data.
Importance A more meaningful approach for microbiome abundance testing is presented. Instead of testing each taxon separately we propose to explore abundance balances among groups of taxa. This approach acknowledges the compositional nature of microbiome data.