mTAGs：基于核糖体 RNA 基因简并共识参考序列的分类剖析。

mTAGs: taxonomic profiling using degenerate consensus reference sequences of ribosomal RNA genes.

机构信息

Department of Biology, Institute of Microbiology and Swiss Institute of Bioinformatics, ETH Zürich, 8093 Zürich, Switzerland.

Department of Gut Microbes and Health, Quadram Institute Bioscience, NR4 7UQ Norwich, UK.

出版信息

Bioinformatics. 2021 Dec 22;38(1):270-272. doi: 10.1093/bioinformatics/btab465.

DOI:10.1093/bioinformatics/btab465

PMID:34260698

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC8696115/

Abstract

UNLABELLED

Profiling the taxonomic composition of microbial communities commonly involves the classification of ribosomal RNA gene fragments. As a trade-off to maintain high classification accuracy, existing tools are typically limited to the genus level. Here, we present mTAGs, a taxonomic profiling tool that implements the alignment of metagenomic sequencing reads to degenerate consensus reference sequences of small subunit ribosomal RNA genes. It uses DNA fragments, that is, paired-end sequencing reads, as count units and provides relative abundance profiles at multiple taxonomic ranks, including operational taxonomic units based on a 97% sequence identity cutoff. At the genus rank, mTAGs outperformed other tools across several metrics, such as the F1 score by >11% across data from different environments, and achieved competitive (F1 score) or better results (Bray-Curtis dissimilarity) at the sub-genus level.

AVAILABILITY AND IMPLEMENTATION

The software tool mTAGs is implemented in Python. The source code and binaries are freely available (https://github.com/SushiLab/mTAGs). The data underlying this article are available in Zenodo, at https://doi.org/10.5281/zenodo.4352762.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

摘要

未加标签

微生物群落的分类组成通常涉及到核糖体 RNA 基因片段的分类。为了在保持高分类准确性的同时进行权衡，现有的工具通常仅限于属级水平。在这里，我们提出了 mTAGs，这是一种分类分析工具，它实现了将宏基因组测序reads 与小亚基核糖体 RNA 基因的简并共识参考序列进行比对。它使用 DNA 片段，即成对测序 reads 作为计数单位，并提供多个分类等级的相对丰度分布情况，包括基于 97%序列同一性截断的操作分类单位。在属级水平上，mTAGs 在多个指标上都优于其他工具，例如在来自不同环境的数据中，F1 分数高出 11%，在亚属水平上也实现了有竞争力的（F1 分数）或更好的结果（Bray-Curtis 不相似性）。