Evaluation of variant calling for cpn60 barcode sequence-based microbiome profiling.

Sarah J Vancuren,Janet E Hill,Scott J Dos Santos,Qunfeng Dong

doi:10.1371/journal.pone.0235682

Abstract

Amplification and sequencing of conserved genetic barcodes such as the cpn60 gene is a common approach to determining the taxonomic composition of microbiomes. Exact sequence variant calling has been proposed as an alternative to previously established methods for aggregation of sequence reads into operational taxonomic units (OTU). We investigated the utility of variant calling for cpn60 barcode sequences and determined the minimum sequence length required to provide species-level resolution. Sequence data from the 5´ region of the cpn60 barcode amplified from the human vaginal microbiome (n = 45), and a mock community were used to compare variant calling to de novo assembly of reads, and mapping to a reference sequence database in terms of number of OTU formed, and overall community composition. Variant calling resulted in microbiome profiles that were consistent in apparent composition to those generated with the other methods but with significant logistical advantages. Variant calling is rapid, achieves high resolution of taxa, and does not require reference sequence data. Our results further demonstrate that 150 bp from the 5´ end of the cpn60 barcode sequence is sufficient to provide species-level resolution of microbiota.

Highlights

Microbiome profiling is the process of determining which organisms are present in an environment, and their relative abundances
It was desirable to first determine an optimal truncation length to use with the cpn60 UT sequences, given input reads of up to 400 bp and the trade-off between minimum length requirement, and proportion of reads retained in variant calling
To investigate the optimal truncation length for cpn60 variant calling with DADA2 without the confounding effect of diminishing read numbers retained as the truncation length requirement increases, we generated a single set of input reads from the vaginal microbiome sequence data that were all at least 300 bp in length after quality filtering

Summary

Introduction

Microbiome profiling is the process of determining which organisms are present in an environment, and their relative abundances. Profiling can be achieved by a “metabarcoding” approach involving targeted PCR and sequencing a genetic barcode: a conserved gene that is shared by many species and can be used to distinguish one from another [1]. Barcodes that have been demonstrated to meet the criteria established by the International Barcode of Life project [2] include the cpn gene in bacteria [3], ITS in fungi [4] and cytochrome c oxidase subunit 1 (COI) in animals [2]. Barcodes are important tools for distinguishing species when phenotypic differences are not conclusive; this is especially useful for prokaryotes. Barcode sequences must be universally conserved, so that a wide range of species can be distinguished by the sequence.

Objectives

Methods

Results