PM-profiler: a high-resolution and fast tool for taxonomy annotation of amplicon-based microbiome

Abstract

ABSTRACT Amplicon sequencing stands as a cornerstone in microbiome profiling, yet concerns persist regarding its resolution and accuracy. The enhancement of reference databases and annotations marks a new era for 16S rRNA-based profiling. Capitalizing on this potential, we introduce PM-profiler, a novel tool for profiling amplicon short reads. PM-profiler is implemented by C++-based advanced algorithms, such as pre-allocated hash for reference construction, hybrid and dynamic short-read matching, big-data-guided dual-mode hierarchical taxonomy annotation strategy, and full-procedure parallel computing. This tool delivers species-level resolution and ultrafast speed for large-scale microbiomes, surpassing alignment-based approaches and the Naïve-Bayesian model. Furthermore, recognizing the global uneven distribution of microbes, we delineate optimal annotation strategies for each sampling habitat based on microbial patterns over 270,000 microbiomes. Integrated with the established workflow of Parallel-Meta Suite and the latest curated reference databases, this endeavor offers a swift and dependable solution for high-precision microbiome surveys. IMPORTANCE Our study introduces PM-profiler, a new tool that deciphers the complexity of microbial communities. With advanced algorithms, flexible annotation strategies, and well-organized big-data, PM-profiler provides a faster and more accurate way to study on microbiomes, paving the way for discoveries that could improve our understanding of microbiomes and their impact on the world.