Overview

phyloBARCODER is a web-based tool for species identification of metabarcoding DNA sequences through phylogenetic tree estimation. It identifies candidate species and selects sequences belonging to the same species from read data. The analysis pages are organized into Animals (mitochondrial genes), Three Domains (SSU rRNA), Plants (plant rbcL), and Fungi (ITS1, ITS2, and LSU rRNA).

Users upload sequences obtained from environmental DNA or metabarcoding samples, referred to here as anonymous sequences. Custom reference sequences can also be supplied as a database. Even when reference sequences have the same BLAST identity to a query, phylogenetic analysis may distinguish a more closely related sequence (result674_Tree_vs_similarity.zip).

New analyses: Three Domains, Plants, and Fungi

Select an analysis page for the marker region of interest. These tools are a public beta. Validation results and bug reports are welcome.

Three Domains: SSU rRNA (PR2 / SILVA)

The Three Domains analysis page primarily targets eukaryotic 18S rRNA sequences. Under Reference database, users can select PR2 v5.1.1, SILVA SSU 144 Parc, or SILVA SSU 144 Ref NR99. PR2 focuses on protists and also includes animals, fungi, and plants. SILVA contains bacterial and archaeal 16S rRNA as well as eukaryotic 18S rRNA. The full PR2 database also includes bacterial, archaeal, and organellar SSU sequences.

Parc provides a broad collection of sequences. Ref NR99 reduces redundancy at 99% sequence similarity after quality and sequence-length filtering. Use reference sequences from the same marker region as the input sequences. Buttons are available to load 18S V4 environmental sequences (PR2) and bacterial or archaeal 16S V4–V5 environmental sequences.

Plants: chloroplast rbcL

The Plants analysis page targets plant chloroplast rbcL sequences. Its reference database combines the Bell reference library with fern data from FTOL and Kuo. This database does not cover all plant species worldwide. If needed, add reference sequences from the focal plants and their relatives to User DB.

Buttons load sequences from UK airborne pollen (Brennan 2019) and pollen collected by bees in Florida (Bell 2017). The rbcL reference button provides an example using Pinus reference sequences.

Fungi: ITS1, ITS2, and LSU rRNA

On the Fungi analysis page, select ITS1, ITS2, or LSU rRNA (28S) under Marker region. ITS1 and ITS2 searches use reference sequences for the selected region from UNITE v10.0 (19 February 2025). For LSU, select SILVA LSU 138.2 Parc or Ref NR99. The SILVA LSU databases also include organisms other than fungi.

Match Marker region to the input sequences. ITS1 and ITS2 are different regions and should not be mixed in an analysis. The ITS loading buttons provide demonstrations using reference sequences. LSU includes a 27-OTU environmental DNA example from decaying wood (Shirouzu et al. 2020), allowing users to compare trees and sequence matches obtained with Parc and Ref NR99.

Common workflow

Paste sequences in FASTA format into Anonymous sequences or load them from a file. Under Number of queries, specify how many sequences from the beginning of the input to use as queries. Select the reference database, the number of hits and E-value for each database, and click SUBMIT. Custom reference sequences can be added to User DB. Inspect the resulting tree and alignment to evaluate species identifications based on relationships to reference sequences. Sections (A) and (B) below illustrate the original workflow using animal mitochondrial genes.

(A) Species Identification

Here, we use 12S rRNA eDNA sequences, as an example of anonymous sequences for uploading. Yu et al. (2022) amplified those sequences by using MiFish primers.

The user copy & pastes anonymous eDNA sequences obtained from the link, Fish 12S, to the text box. To identify species names for 2 OTUs simultaneously, Number of queries is set as First 2 sequences.

Number of queries: This box has a maximum of “First 10 sequences.” When “First 2 sequences” are selected, only the first 2 are used as queries for BLAST searches. However, all anonymous sequences including 2 query sequences are converted to “Anonymous DB.” Therefore, anonymous sequences having close matches with the first 2 sequences are included in the multiple sequence alignment. The remaining anonymous sequences (not selected by BLAST) are not included in the alignment.

-num_alignments or -evalue options: For all databases, those options clarify species identification by adjusting sequence members included in resultant alignments.
- For Anonymous DB, these numbers increase or decrease the number of BLAST hits in order to clarify anonymous sequences belonging to the same groups of query sequences.
-For Species DB (Pre-installed DB), those numbers can identify appropriate root sequences for focal species/groups in resultant phylogenetic trees, and for Haplotype DB, they delineate species/group or population groups clearer with different sequences of the same reference species.
- For User DB, those numbers increase or decrease related sequences of focal species and relatives.
- If “- num_alignments” is set as “0 (not used)”, the database is not used.

The results can be shown by clicking the link next to Status > Finished.

Above results can be downloaded (result1192_phyloBARCODER.zip). With reference to the estimated phylogenetic tree, the user needs to evaluate species identifications for the queries by sight.

Also, the user can evaluate the species identification from the alignment.

As a species list, phyloBARCODER automatically produces species identifications for the user-defined queries. The species identification and classification* for each query are produced from BLAST hits derived from Pre-installed DB and are saved in the “taxon_assignment_tree.csv” file. For the OTU_6 species name, not only Scomber japonicus but also S. australasicus and S. colias are candidates. To further narrow down the species name, distribution of each species can be considered.
*Those are not produced for BLAST hits from User DB (the custom user reference data).

Example: Fish ASV analysis

Anonymous sequences

KS1815-B06-0m_ASV.fasta.txt

Raw data

KS1815-B06-0m_S50_R1_001.fastq.gz

 

KS1815-B06-0m_S50_R2_001.fastq.gz

 

(Yu et al. 2021)

Example: Copepod analysis

Reference sequence

Metridia_pacifica_lucens_MZGdb_Selected.txt

 

(MetaZooGene Atlas & Database, 11 Oct 2023)

(B) Sequence Extraction

For reference sequences of 12S rRNA gene, we count the number of Scomber species.

Select or enter the following parameters:

Pre-installed DB

Species

Gene

12S (srRNA)

Classification

Scomber

Scomber 12S rRNA sequences are found for all 4 known Scomber species. This indicates that misidentification is unlikely due to absence of reference sequences of Scomber, if the query is known to be included in Scomber before the phyloBARCODER identification.

(C) BLAST species identification

phyloBARCODER provides online BLAST searches against MIDORI2, covering 15 metazoan mitochondrial genes; MiFish_DB v50, covering the fish mitochondrial 12S rRNA gene; or PR2 v5.1.1, covering SSU rRNA sequences, mainly eukaryotic 18S, with bacterial, archaeal and organellar sequences also included. Users can identify candidate species from eDNA and metabarcoding sequences in FASTA format and download up to three BLAST hits per query as a CSV file.

Citation

Inoue J. et al. phyloBARCODER: A web tool for phylogenetic classification of eukaryote metabarcodes using custom reference databases. Molecular Biology and Evolution 2024, 42: msae111. doi: 10.1093/molbev/msae111

Dependencies

Similarity search

BLAST+ (blastn 2.7.1)

Alignment

MAFFT v7.490; trimAl 1.2rev59

Tree search

ape in R, Version 5.6.2

Pre-installed database

MIDORI2 longest (species) and uniq (haplotype)

History