micRoclean: an R package for decontaminating low-biomass 16S-rRNA microbiome data.

Area of Science:

• Microbiology
• Bioinformatics

Background:

• Cross-contamination and environmental contamination are significant challenges in 16S-rRNA microbiome studies, especially in low-biomass samples.
• Existing decontamination tools lack consensus on selection criteria and methods to quantify the impact of contaminant removal, risking over-filtering.

Purpose of the Study:

• To introduce micRoclean, an open-source R package designed to address the gaps in microbiome data decontamination.
• To provide researchers with two distinct decontamination pipelines and guidance for selecting the appropriate one based on study design and goals.

Main Methods:

• Developed micRoclean, an R package integrating and expanding upon existing decontamination and analysis tools.
• Implemented a filtering loss statistic to quantify the impact of decontamination on the data's covariance structure.
• Validated micRoclean using example data and a multi-batch simulated microbiome sample.

Main Results:

• micRoclean effectively and intuitively decontaminates microbiome data.
• The package provides clear guidance for selecting appropriate decontamination pipelines.
• The filtering loss statistic quantifies the impact of decontamination on data covariance.
• micRoclean matches or outperforms existing tools in simulated multi-batch samples.

Conclusions:

• micRoclean offers a comprehensive and user-friendly solution for microbiome data decontamination.
• The package aids researchers in improving the accuracy of low-biomass microbiome studies by providing reliable decontamination methods.
• micRoclean is freely available on GitHub, promoting accessibility and further development in the field.