Department of Microbiology and Immunology, Institute for Genome Sciences, University of Maryland School of Medicine, Baltimore, MD, 21201, USA.
BMC Bioinformatics. 2024 Sep 27;25(1):313. doi: 10.1186/s12859-024-05930-3.
Clustering of sequences into operational taxonomic units (OTUs) and denoising methods are a mainstream stopgap to taxonomically classifying large numbers of 16S rRNA gene sequences. Environment-specific reference databases generally yield optimal taxonomic assignment.
We developed SpeciateIT, a novel taxonomic classification tool which rapidly and accurately classifies individual amplicon sequences ( https://github.com/Ravel-Laboratory/speciateIT ). We also present vSpeciateDB, a custom reference database for the taxonomic classification of 16S rRNA gene amplicon sequences from vaginal microbiota. We show that SpeciateIT requires minimal computational resources relative to other algorithms and, when combined with vSpeciateDB, affords accurate species level classification in an environment-specific manner.
Herein, two resources with new and practical importance are described. The novel classification algorithm, SpeciateIT, is based on 7th order Markov chain models and allows for fast and accurate per-sequence taxonomic assignments (as little as 10 min for 10 sequences). vSpeciateDB, a meticulously tailored reference database, stands as a vital and pragmatic contribution. Its significance lies in the superiority of this environment-specific database to provide more species-resolution over its universal counterparts.
将序列聚类为操作分类单元（OTUs）和去噪方法是对大量 16S rRNA 基因序列进行分类的主流权宜之计。特定于环境的参考数据库通常可实现最佳的分类分配。
我们开发了 SpeciateIT，这是一种新颖的分类工具，可快速准确地对单个扩增子序列进行分类（https://github.com/Ravel-Laboratory/speciateIT）。我们还介绍了 vSpeciateDB，这是一个用于阴道微生物群 16S rRNA 基因扩增子序列分类的自定义参考数据库。我们表明，SpeciateIT 相对于其他算法所需的计算资源最少，并且与 vSpeciateDB 结合使用时，可以以特定于环境的方式提供准确的物种水平分类。
本文描述了两个具有新的实用重要性的资源。新的分类算法 SpeciateIT 基于 7 阶马尔可夫链模型，可实现快速准确的序列分类（10 个序列的分类时间短至 10 分钟）。vSpeciateDB 是一个精心制作的参考数据库，是一个重要的实用贡献。其重要性在于该特定于环境的数据库相对于通用数据库在提供更多物种分辨率方面的优势。
