Robust regression with compositional covariates

Many biological high-throughput datasets, such as targeted amplicon-based and metagenomic sequencing data, are compositional. A common exploratory data analysis task is to infer robust statistical associations between high-dimensional microbial compositions habitat- or host-related covariates. To address this, a general regression framework RobRegCC (Robust Regression with Compositional Covariates) proposed, which extends the linear log-contrast model by mean shift formulation for capturing outliers. includes sparsity-promoting convex non-convex penalties parsimonious estimation, data-driven initialization procedure, novel cross-validation selection scheme. The procedure implemented in R package robregcc. Extensive simulation studies show RobRegCC's ability perform simultaneous sparse outlier detection over wide range of settings. demonstrate seamless applicability workflow real gut microbiome dataset from HIV patients analyzed set species host immune response soluble CD14 measurements inferred.