Alignment-Free Viral Sequence Classification at Scale
Author Information
Author(s): van Zyl Daniel J., Dunaiski Marcel, Tegally Houriiyah, Baxter Cheryl, The INFORM Africa research study group, de Oliveira Tulio, Xavier Joicymara S.
Primary Institution: Cold Spring Harbor Laboratory
Hypothesis
This study evaluates alignment-free methods as scalable and rapid alternatives for viral sequence classification.
Conclusion
The study demonstrates that alignment-free methods can classify viral sequences quickly and accurately using less computational power.
Supporting Evidence
- The primary dataset comprises 297,186 SARS-CoV-2 nucleotide sequences, categorized into 3502 distinct lineages.
- The AF classifiers achieved 97.8% accuracy on the SARS-CoV-2 test set.
- The classifiers achieved 99.8% accuracy on dengue test sets.
- The classifiers achieved 89.1% accuracy on HIV test sets.
- AF techniques process data significantly faster than alignment-based methods.
Takeaway
This study shows a new way to quickly and accurately sort viruses using computer methods that don't need to line up their genetic codes.
Methodology
Six established alignment-free techniques were used to extract feature vectors from viral genomes, which were then used to train Random Forest classifiers.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website