A configuration-resolved benchmark of differential abundance analysis methods for human gut 16S rRNA microbiome data
Tools for differential abundance testing of 16S rRNA data are conventionally treated as discrete methods, and benchmarks have accordingly sought to determine which tool performs best. However, each tool offers an array of configurations based on different normalisation, transformation, reference choice, and sensitivity filtering methods, and the specific impact of these configurations on…
A comprehensive evaluation of differential abundance testing methods for 16S rRNA human gut microbiome data reveals that the performance of these tools is heavily influenced by the specific configurations employed. Traditionally, these tools are viewed as discrete options, and benchmarks have focused on identifying the single best tool.
However, each tool encompasses a variety of configurations related to normalization, transformation, reference selection, and sensitivity filtering. The impact of these configurations on tool performance has not been thoroughly analyzed.
The study tested five popular tools (MaAsLin 2, MaAsLin 3, edgeR, ALDEx2, and ANCOM-BC2) across 18 different configurations. These simulated and real-world human gut microbiome data underwent analysis at two taxonomic resolutions and across various design factors. The researchers discovered that configuration accounted for just as much variation in performance as the selection of the tool itself. The ranking of two tools changed depending on the specific settings being compared.
Furthermore, individual parameters appeared as binary switches between two contrasting error regimes instead of continuous adjustments. Identifiable settings within each tool were found to carry greater weight in determining performance. These findings were consistent across different data sources and taxonomic resolutions. The study establishes a framework that recommends matching a suitable tool with its optimal configurations based on the cohort, study design, and feature resolution.
This ensures that differential abundance results are only interpretable if the used configuration is reported.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
