Taxonomic classification of metabarcoding data involves several key steps to accurately identify and assign taxonomy to the sequences obtained from environmental samples. This process is crucial for understanding the biodiversity and ecological dynamics within a given environment. This section provides a general overview of how this classification is achieved.

Alignment to reference sequences

The processed sequences, such as Amplicon Sequence Variants (ASVs) or Operational Taxonomic Units (OTUs), are aligned to reference sequences contained in established databases. This alignment is the first step in determining the identity of the sequences.

Use of relevant bioinformatics Tools

Several bioinformatics tools are employed to compare the ASVs/OTUs to reference sequences. Commonly used tools include:

BLAST: This tool compares nucleotide sequences to sequence databases and calculates the statistical significance of matches.

USEARCH: Known for its efficiency in sequence alignment and clustering, USEARCH is widely used for identifying and classifying sequences (Edgar RC, 2010).

RDP Classifier (Naïve Bayesian): The Ribosomal Database Project (RDP) classifier uses a naïve Bayesian algorithm to assign taxonomic classifications based on the probability of sequence similarity (Wang Q et al, 2007). This method is also implemented with the ‘assignTaxonomy’ function in DADA2.

Confidence in taxonomic assignment

The bioinformatics tools provide probability or bootstrapping values which indicate the confidence level of the taxonomic assignment. Confidence values help in determining the reliability of the assigned taxonomy. For example, a high confidence value means that the assigned taxonomy is very likely to be correct. Assignments are made at the highest taxonomic level with the highest confidence. In cases where high confidence is achieved only at a broader taxonomic level (e.g. kingdom level of Fungi), the assignment will be made at that level rather than a more specific level (e.g. species or genus). Several published articles have reported studies comparing various taxonomic assigners across different datasets (Bokulich et al, 2018; Hleap et al, 2021).

You can move on to the next page to explore the open resources available for taxonomy assignments in biodiversity research.