摘要：Contamination is a ubiquitous problem in microbiome research and can skew results, especially when small amounts of target DNA are available. Nevertheless, no clear solution has emerged for removing microbial contamination. To address this problem, we developed the R package microDecon ( https://github.com/donaldtmcknight/microDecon ), which uses the proportions of contaminant operational taxonomic units (OTUs) or amplicon sequence variants (ASVs) in blank samples to systematically identify and remove contaminant reads from metabarcoding data sets. We rigorously tested microDecon using a series of computer simulations and a sequencing experiment. We also compared it to the common practice of simply removing all contaminant OTUs/ASVs and other methods for removing contamination. Both the computer simulations and our sequencing data confirmed the utility of microDecon. In our largest simulation (100,000 samples), using microDecon improved the results in 98.1% of samples. Additionally, in the sequencing data and in simulations involving groups, it enabled accurate clustering of groups as well as the detection of previously obscured patterns. It also produced more accurate results than the existing methods for identifying and removing contamination. These results demonstrate that microDecon effectively removes contamination across a broad range of situations. It should, therefore, be widely applicable to microbiome studies, as well as to metabarcoding studies in general.
