Have you ever had to perform taxonomic classification of fungal ITS sequences and were not happy with the results? Well, if you had this problem, I would like to invite you to try our hierarchical taxonomic classifier called HiClass to possibly improve your predictions.
We benchmarked HiTaC against 17 different taxonomic classifiers available in the TAXXI benchmark. Our results show that HiTaC:
- Improved the accuracy and sensitivity upon state-of-the-art classifiers on noisy data;
- Had a high accuracy and sensitivity when predicting sequences known in the database, tying with BTOP;
To simplify your life, we implemented HiClass as a QIIME2 plugin and it can be easily installed with pip install hitac, conda install -c conda-forge -c bioconda hitac or docker pull mirand863/hitac:latest.
You can read the full manuscript on bioRxiv https://doi.org/10.1101/2020.04.24.014852, read the documentation and reproduce the benchmark that are available in our repository https://github.com/mirand863/hitac
We are very open to feedback, so if you find bugs, would like to request new features or pre-trained models with your favorite database, etc., just reach out to Fabio.MalcherMiranda@hpi.de
Best regards,
Fabio