HiTaC

HiTaC is an open-source hierarchical taxonomic classifier for fungal ITS sequences.

Quick links

• Benchmark
• Install standalone version
• Quick start for standalone version
• Install as a QIIME2 plugin
• Quick start for QIIME2 plugin
• Pre-trained models
• Support
• Contributing
• Getting the latest updates
• Citation

Benchmark

HiTaC was thoroughly evaluated with the TAXXI benchmark, consistently achieving higher accuracy and sensitivity as evidenced in the figures below.

For reproducibility, a Snakemake pipeline was created. Instructions on how to run it and source code are available at https://github.com/mirand863/hitac/tree/main/benchmark.

Install standalone version

Option 1: Conda

HiTaC can be easily installed in a new conda environment by running the following command:

conda create -n hitac -c bioconda hitac

Afterward, the new conda environment created can be activated with:

conda activate hitac

For conda installation instructions, we refer the reader to Conda's user guide.

Option 2: Pip

Alternatively, HiTaC can be installed with pip by running:

pip install hitac

Option 3: Docker

Lastly, HiTaC can be downloaded as a docker image:

docker pull mirand863/hitac_standalone:latest

The downloaded image can then be started with:

docker run -it mirand863/hitac_standalone:latest /bin/bash

Quick start for standalone version

For an interactive tutorial, we refer the reader to our Google Colabs notebook.

To see the usage run [command] --help if you want further help with a specific command.

usage: hitac-fit [-h] --reference REFERENCE [--kmer KMER] [--threads THREADS] --classifier CLASSIFIER

Fit hierarchical classifier

optional arguments:
  -h, --help            show this help message and exit
  --reference REFERENCE
                        Input FASTA file with reference sequence(s) to train model
  --kmer KMER           K-mer size for feature extraction [default: 6]
  --threads THREADS     Number of threads to train in parallel [default: all]
  --classifier CLASSIFIER
                        Path to store trained hierarchical classifier

Input Files

HiTaC accepts reference and query files in FASTA format. The reference file must have the taxonomies annotated as follows:

>EU272527;tax=d:Fungi,p:Ascomycota,c:Eurotiomycetes,o:Eurotiales,f:Trichocomaceae,g:Paecilomyces,s:Paecilomyces_sinensis;
CCGAGTGAGGGTCCCACGAGGCCCAACCTCCCATCCGTGTTGAACTACACCTGTTGCTTCGGCGGGCCCGCCGTGGTTCA
CGCCCGGCCGCCGGGGGGCCTTGTGCTCCCGGGCCCGCGCCCGCCGAAGACCCCTCGAACGCTGCCCTGAAGGTTGCCGT
CTGAGTATAAAATCAATCATTAAAACTTTCAACAACGGATCTCTTGGTTCCGGCATCGATGAAGAACGCAGCGAAATGCG
ATAAGTAATGTGAATTGCAGAATTCCGTGAATCATCGAATCTTTGAACGCACATTGCGCCCCCTGGCATTCCGGGGGGCA
TGCCTGTCCGAGCGTCATTGCTAACCCTCCAGCCCGGCTGGTGTGTTGGGTCGACGTCCCCCCCGGGGGACGGGCCCGAA
AGGCAGCGGCGGCGCCGCGTCCGATCCTCGAGCGTATGGGGCTTTGTCACGCGCTCTGGTAGGGTCGGCCGGCTGGCCAG
CCAGCGACCTCACGGTCACCTATTTTTTCTCTTAGG
>L54118;tax=d:Fungi,p:Basidiomycota,c:Agaricomycetes,o:Boletales,f:Suillaceae,g:Suillus,s:Suillus_placidus;
ACGAATTCATAATTCGGCGAGGGGAAAGCGGAGGGTTGTAGCTGGCCTTTTTACCGAGGCACGTGCACGCTCTCTTCCGA
ACTCTCGTCGTATGGGCGCGGGGCGACCCGCGTCTTTCATCCCACCTCTTCGTGTAGAAAGTCTTTGAATGTTTTTACCA
TCATCGAGTCGCGACTTCTAGGAGACGCGATTCTTTGAGACAAAAGTTTATTACAACTTTCAGCAATGGATCTCTTGGCT
CTCGCATCGATGAAGAACGCAGCGAATCGCGATATGTAATGTGAATTGCAGATCTACAGTGAATCATCGAATCTTTGAAC
GCACCTTGCGCTCCTCGGTGTTCCGAGGAGCATGCCTGTTTGAGCGTCAGTAAATTCTCAACCCCTCTCGATTTGCTTCG
AGAGGGCGCTTGGATGGTGGGGGCTGCCGGAGACCTGGATTTATCCCTGGACTCGGGCTCTCCTGAAATGCATCGGCTTG
CGGTCGACTTTCGACTTTGCGCGACAAGGCCTTCGGCGTGATAATGATCGCCGTTCGCCGAAGCGCAGGAATGAACGGTC
CCGCGCCTCTAATCCGTCGACGCTTTCGAGCGTCTTCCTCATTGACGTTTGACCTCAAAT

Training and predicting taxonomies

To train the model and classify, simply run:

hitac-fit \
--reference reference.fasta \
--classifier classifier.pkl

hitac-classify \
--classifier classifier.pkl \
--reads reads.fasta \
--classification classification.tsv

Additionally, a filter can be trained to remove ranks where the predictions might be inaccurate and to compute the confidence score:

hitac fit-filter \
--reference reference.fasta \
--filter filter.pkl

hitac filter \
--filter filter.pkl \
--reads reads.fasta \
--classification classification.tsv \
--filtered-classification filtered_classification.tsv

Output File

HiTaC generates a TSV file for the predictions. The first column in the TSV file contains the identifier of the test sequence and the second column holds the predictions made by HiTaC. For example:

EU254776    d:Fungi,p:Ascomycota,c:Sordariomycetes,o:Diaporthales,f:Valsaceae,g:Cryptosporella,s:Cryptosporella_femoralis
FJ711636    d:Fungi,p:Basidiomycota,c:Agaricomycetes,o:Agaricales,f:Marasmiaceae,g:Armillaria,s:Armillaria_tabescens
UDB016040   d:Fungi,p:Basidiomycota,c:Agaricomycetes,o:Russulales,f:Russulaceae,g:Russula,s:Russula_adusta
GU827310    d:Fungi,p:Ascomycota,c:Lecanoromycetes,o:Lecanorales,f:Ramalinaceae,g:Ramalina,s:Ramalina_conduplicans
JN943699    d:Fungi,p:Ascomycota,c:Lecanoromycetes,o:Lecanorales,f:Parmeliaceae,g:Punctelia,s:Punctelia_caseana

Install as a QIIME2 plugin

Option 1: Conda

HiTaC can also be installed as a QIIME 2 plugin. To install QIIME 2 version 2023.2 in a GNU/Linux machine, run:

wget https://data.qiime2.org/distro/core/qiime2-2023.2-py38-linux-conda.yml
conda env create -n hitac --file qiime2-2023.2-py38-linux-conda.yml
# OPTIONAL CLEANUP
rm qiime2-2023.2-py38-linux-conda.yml

Note: Instructions on how to install on Windows and macOS are available at QIIME 2 docs.

Afterward, the new conda environment created in the last step can be activated and HiTaC can be installed:

conda activate hitac
conda install -c conda-forge -c bioconda hitac

For conda installation instructions, we refer the reader to Conda's user guide.

Option 2: Pip

Alternatively, HiTaC can be installed with pip in an environment where QIIME 2 was previously installed:

pip install hitac

Option 3: Docker

Lastly, HiTaC and all its dependencies can be downloaded as a docker image:

docker pull mirand863/hitac_qiime:latest

The downloaded image can then be started with:

docker run -it mirand863/hitac_qiime:latest /bin/bash

Quick start for QIIME2 plugin

For an interactive tutorial, we refer the reader to our Google Colabs notebook.

To see the usage run qiime hitac --help or qiime hitac [command] --help if you want further help with a specific command.

Usage: qiime hitac [OPTIONS] COMMAND [ARGS]...

  Description: This QIIME 2 plugin wraps HiTaC for hierarchical taxonomic
  classification.

  Plugin website: https://gitlab.com/dacs-hpi/hitac

  Getting user support: Please post to the QIIME 2 forum for help with this
  plugin: https://forum.qiime2.org

Options:
  --version    Show the version and exit.
  --citations  Show citations and exit.
  --help       Show this message and exit.

Commands:
  classify    Hierarchical classification with HiTaC's pre-fitted model
  filter      Hierarchical classification filtering with HiTaC's pre-fitted
              model

  fit         Train HiTaC's hierarchical classifier
  fit-filter  Train HiTaC's hierarchical filter

Input Files

HiTaC accepts taxonomy in TSV format and training and test files in FASTA format. All these files must be previously imported by QIIME 2, for example:

qiime tools import \
--input-path query.fasta \
--output-path query.qza \
--type 'FeatureData[Sequence]'

qiime tools import \
--input-path reference.fasta \
--output-path reference.qza \
--type 'FeatureData[Sequence]'

qiime tools import \
--type 'FeatureData[Taxonomy]' \
--input-format HeaderlessTSVTaxonomyFormat \
--input-path taxonomy.txt \
--output-path taxonomy.qza

Training and predicting taxonomies

To train the model and classify, simply run:

qiime hitac fit \
--i-reference-reads reference.qza \
--i-reference-taxonomy taxonomy.qza \
--o-classifier classifier.qza

qiime hitac classify \
--i-classifier classifier.qza \
--i-reads query.qza \
--o-classification classification.qza

Additionally, a filter can be trained to remove ranks where the predictions might be inaccurate and to compute the confidence score:

qiime hitac fit-filter \
--i-reference-reads reference.qza \
--i-reference-taxonomy taxonomy.qza \
--o-filter filter.qza

qiime hitac filter \
--i-filter filter.qza \
--i-reads query.qza \
--i-classification classification.qza \
--o-filtered-classification filtered_classification.qza

Output File

The predictions can be exported from QIIME 2 to a TSV file:

qiime tools export \
--input-path classification.qza \
--output-path output_dir

or alternativelly if the filter was used:

qiime tools export \
--input-path filter_output.qza \
--output-path output_dir

The first column in the TSV file contains the identifier of the test sequence, while the second column holds the predictions made by HiTaC and the third column is the prediction probability if the filter was applied. For example:

Feature ID  Taxon   Confidence
EU254776    d__Fungi; p__Ascomycota; c__Sordariomycetes; o__Diaporthales; f__Valsaceae; g__Cryptosporella   -1
FJ711636    d__Fungi; p__Basidiomycota; c__Agaricomycetes; o__Agaricales; f__Marasmiaceae; g__Armillaria    -1
UDB016040   d__Fungi; p__Basidiomycota; c__Agaricomycetes; o__Russulales; f__Russulaceae; g__Russula    -1
GU827310    d__Fungi; p__Ascomycota; c__Lecanoromycetes; o__Lecanorales; f__Ramalinaceae; g__Ramalina   -1
JN943699    d__Fungi; p__Ascomycota; c__Lecanoromycetes; o__Lecanorales; f__Parmeliaceae; g__Punctelia  -1

Pre-trained models

In order to speed up the process for users, we provide pre-trained models on the public database UNITE. Some of these pre-trained models contain all eukaryotic ITS sequences available on UNITE, which enable detection and removal of nonfungal sequences mistakenly amplified by polymerase chain reaction (PCR). Furthermore, HiTaC uses the unique species hypotheses identifiers provided in the database UNITE as the last taxonomic level in the hierarchy during training and reports them to the user, which increases taxonomic reproducibility.

• Pre-trained QIIME2 models for HiTaC (UNITE all eukaryotes)
• Pre-trained QIIME2 models for HiTaC (UNITE only fungi)
• [Pre-trained QIIME2 models for HiTaC_Filter (UNITE all eukaryotes)]()
• [Pre-trained QIIME2 models for HiTaC_Filter (UNITE only fungi)]()
• Pre-trained standalone models for HiTaC (UNITE eukaryotes global)
• Pre-trained standalone models for HiTaC (UNITE eukaryotes refs)
• Pre-trained standalone models for HiTaC (UNITE only fungi)
• [Pre-trained standalone models for HiTaC_Filter (UNITE all eukaryotes)]()
• [Pre-trained standalone models for HiTaC_Filter (UNITE only fungi)]()

Support

If you run into any problems or issues, please create a GitLab issue and we will try our best to help.

We strive to provide good support through our issue tracker on GitLab. However, if you'd like to receive private support with:

• Phone / video calls to discuss your specific use case and get recommendations
• Private discussions over Slack or Mattermost

Please reach out to fabio.malchermiranda@hpi.de.

Contributing

We are a small team on a mission to improve ITS taxonomic classification, and we will take all the help we can get! If you would like to get involved, here is information on contribution guidelines and how to test the code locally.

You can contribute in multiple ways, e.g., reporting bugs, writing or translating documentation, reviewing or refactoring code, requesting or implementing new features, etc.

Getting the latest updates

If you'd like to get updates when we release new versions, please click on the notification button on the top and select "Watch". GitLab will then send you notifications along with a changelog with each new release.

Citation

If you use HiTaC, please cite:

Miranda, Fábio M., et al. "HiTaC: Hierarchical Taxonomic Classification of Fungal ITS Sequences." bioRxiv (2020).

@article{miranda2020hitac,
  title={HiTaC: Hierarchical Taxonomic Classification of Fungal ITS Sequences},
  author={Miranda, F{\'a}bio M and Azevedo, Vasco AC and Renard, Bernhard Y and Piro, Vitor C and Ramos, Rommel TJ},
  journal={bioRxiv},
  year={2020},
  publisher={Cold Spring Harbor Laboratory}
}