Biology subjects

van Zyl, D. J.

Publications and source records attributed to van Zyl, D. J..

3 recordsLinked to original sources

bioRxiv2026

Anniemap: Vector Search for Viral Short Read Alignment

Background: The process of aligning sequencing reads to a reference genome is a foundational step in genomic analysis, underpinning tasks from variant detection to pathogen surveillance. In viral genomics, however, this problem becomes substantially more challenging: viral sequences are often present at low abundance within host-dominated samples and can differ markedly from available references due to rapid mutation and population heterogeneity. These characteristics reduce the effectiveness of conventional seed-and-extend aligners, which typically rely on long exact or near-exact matches to anchor alignments. Even modest sequence divergence or sequencing errors can disrupt such seeds, particularly for short reads, leading to missed alignments. The central challenge in this setting is maintaining robust alignment under high divergence without sacrificing efficiency. Results: We introduce Anniemap, a vector search based approach to viral short-read sequence alignment. Anniemap represents reads and reference sequences as binary vectors and performs approximate nearest-neighbour search using Facebook AI Similarity Search (FAISS) to efficiently identify candidate mappings. Anniemap was compared with the well-established alignment tools Bowtie2 and BWA-MEM2 across a diverse set of viral genomes and read lengths using both simulated and real sequencing data. Anniemap achieved higher sensitivity and throughput in almost all evaluated scenarios, with the most substantial improvements in sensitivity observed for highly divergent genomes, such as Hepatitis C virus (HCV) and Human Immunodeficiency Virus (HIV). Conclusions; By measuring vector similarity rather than relying on long exact seed matches, Anniemap provides greater robustness to sequencing errors and genomic mutations. This property is particularly advantageous for viral genomes, where substantial sequence divergence is common. Further work is required to efficiently extend vector-based search for read alignment beyond viral genomes.

genomics↗

bioRxiv2025

Craft: A Machine Learning Approach to Dengue Subtyping

MotivationThe dengue virus poses a major global health threat, with nearly 390 million infections annually. A recently proposed hierarchical dengue nomenclature system enhances spatial resolution by defining major and minor lineages within genotypes, aiding efforts to track viral evolution. While current subtyping tools - Genome Detective, GLUE, and NextClade - rely on computationally intensive sequence alignment and phylogenetic inference, machine learning presents a promising alternative for achieving accurate and rapid classification. ResultsWe present Craft (Chaos Random Forest), a machine learning framework for dengue subtyping. We demonstrate that Craft is capable of faster classification speeds while matching or surpassing the accuracy of existing tools. Craft achieves 99.5% accuracy on a hold-out test set and processes over 140 000 sequences per minute. Notably, Craft maintains remarkably high accuracy even when classifying sequence segments as short as 700 nucleotides. Contactdanielvanzyl@sun.ac.za Supplementary informationA supplemental table acknowledging the authors of the GISAID dengue sequences is available at Bioinformatics online.

genomics↗

bioRxiv2024

Alignment Free Viral Sequence Classification

BackgroundThe rapid increase in nucleotide sequence data generated by next-generation sequencing (NGS) technologies demands efficient computational tools for sequence comparison. Alignment-based methods, such as BLAST, are increasingly overwhelmed by the scale of contemporary datasets due to their high computational demands for classification. This study evaluates alignment-free (AF) methods as scalable and rapid alternatives for viral sequence classification, focusing on identifying techniques that maintain high accuracy and efficiency when applied to extremely large datasets. ResultsWe employed six established AF techniques to extract feature vectors from viral genomes, which were subsequently used to train Random Forest classifiers. Our primary dataset comprises 297,186 SARS-CoV-2 nucleotide sequences, categorized into 3502 distinct lineages. Furthermore, we validated our models using dengue and HIV sequences to demonstrate robustness across different viral datasets. Our AF classifiers achieved 97.8% accuracy on the SARS-CoV-2 test set, and 99.8% and 89.1% accuracy on dengue and HIV test sets, respectively. ConclusionDespite the high-class dimensionality, we show that word-based AF methods effectively represent viral sequences. Our study highlights the practical advantages of AF techniques, including significantly faster processing compared to alignment-based methods and the ability to classify sequences using modest computational resources.

genomics↗