Fast and inexpensive visualization of genome collection at scale.
- 1TCS Research, Tata Consultancy Services (TCS) Ltd, Hyderabad, India.
- 2TCS Research, Tata Consultancy Services (TCS) Ltd, Hyderabad, India. [email protected].
- 3TCS Research, Tata Consultancy Services (TCS) Ltd, Hyderabad, India. [email protected].
Abstract
Background
Investigating genetic progression and lineage using sequence data is a fundamental focus within the realm of viral genome research. Given the substantial increase in publicly accessible genome sequences, tools designed for extensive data analytics, particularly insightful visualizations play an important role in elucidating the spatio-temporal evolution as well as the diversity of viral lineages. Existing genome data analysis methods frequently rely on multiple sequence alignment (MSA) and phylogeny computations, making integration of new sequences computationally expensive. Such methods are often metadata agnostic and typically utilize a global distance metric between candidate sequences. We present BOVIZ, a fast and inexpensive method for rich visualization of large-scale genome collections. BOVIZ leverages variant-level features from candidate sequences and employs a novel Bag-Of-Variants (BOV) method that transforms high-dimensional variant feature vectors into a 2D-space, effectively capturing sequence (dis)similarities. BOVIZ circumvents the need for expensive MSA and phylogeny analyses, enabling efficient integration of novel sequences and faster visualization of large-scale sequences on a standard desktop. Furthermore, BOVIZ supports a variety of filters, including metadata filters and selection of genomic regions of interest for effectively capturing localized similarities and dissimilarities on-the-fly. We perform extensive benchmarking of BOVIZ with state-of-the-art visualization methods using the SARS-CoV-2 viral genome dataset.Results
BOVIZ consistently outperformed these methods in capturing inter-sequence similarity and diversity, offering superior visualization of spatio-temporal and clade-level evolution. It is able to effectively distinguish SARS-CoV-2 lineages into distinct clusters, visualize the evolution of specific sublineages like Omicron and their variant patterns, as well as visualize potential sub-cluster within the SARS-CoV-2 Delta variant.Conclusion
BOVIZ is an effective visualization tool for genome surveillance, complementing resource-intensive deep analytics by providing key insights about the underlying genome landscape in a fast and scalable manner.
Related
- The Cat Gut Microbial Genome Collection reveals global structure of the feline gut microbiome
- Population-scale chemical response revealed by a barcoded yeast collection
- Provenance-Associated Confounding in Cross-Collection Comparisons of Urease Inhibitors: A Quantified Reference Scale and a Domain-Aware Alternative
- The Dual-Ribozyme RNA Ladder and RNA-Calibrated ssDNA Marker System for Inexpensive RNA Size Standards
- A genome-scale CRISPRi perturbation atlas of human induced pluripotent stem cells
BibTeX
@article{kotte2026fast,
title = {{Fast and inexpensive visualization of genome collection at scale.}},
author = {Sujatha Kotte and Naina Tiwari and Kavya Vaddadi and Vangala Govindakrishnan Saipradeep and Thomas Joseph and Aditya Rao and Rajgopal Srinivasan and Naveen Sivadasan},
journal = {BMC bioinformatics},
year = {2026},
month = jul,
doi = {10.1186/s12859-026-06498-w},
url = {https://europepmc.org/article/MED/42894138},
note = {CC BY},
}
RIS
TY - JOUR
TI - Fast and inexpensive visualization of genome collection at scale.
AU - Sujatha Kotte
AU - Naina Tiwari
AU - Kavya Vaddadi
AU - Vangala Govindakrishnan Saipradeep
AU - Thomas Joseph
AU - Aditya Rao
AU - Rajgopal Srinivasan
AU - Naveen Sivadasan
PY - 2026
DA - 2026/07/01
JO - BMC bioinformatics
DO - 10.1186/s12859-026-06498-w
UR - https://europepmc.org/article/MED/42894138
AB - Background Investigating genetic progression and lineage using sequence data is a fundamental focus within the realm of viral genome research. Given the substantial increase in publicly accessible genome sequences, tools designed for extensive data analytics, particularly insightful visualizations play an important role in elucidating the spatio-temporal evolution as well as the diversity of viral lineages. Existing genome data analysis methods frequently rely on multiple sequence alignment (MSA) and phylogeny computations, making integration of new sequences computationally expensive. Such methods are often metadata agnostic and typically utilize a global distance metric between candidate sequences. We present BOVIZ, a fast and inexpensive method for rich visualization of large-scale genome collections. BOVIZ leverages variant-level features from candidate sequences and employs a novel Bag-Of-Variants (BOV) method that transforms high-dimensional variant feature vectors into a 2D-space, effectively capturing sequence (dis)similarities. BOVIZ circumvents the need for expensive MSA and phylogeny analyses, enabling efficient integration of novel sequences and faster visualization of large-scale sequences on a standard desktop. Furthermore, BOVIZ supports a variety of filters, including metadata filters and selection of genomic regions of interest for effectively capturing localized similarities and dissimilarities on-the-fly. We perform extensive benchmarking of BOVIZ with state-of-the-art visualization methods using the SARS-CoV-2 viral genome dataset.Results BOVIZ consistently outperformed these methods in capturing inter-sequence similarity and diversity, offering superior visualization of spatio-temporal and clade-level evolution. It is able to effectively distinguish SARS-CoV-2 lineages into distinct clusters, visualize the evolution of specific sublineages like Omicron and their variant patterns, as well as visualize potential sub-cluster within the SARS-CoV-2 Delta variant.Conclusion BOVIZ is an effective visualization tool for genome surveillance, complementing resource-intensive deep analytics by providing key insights about the underlying genome landscape in a fast and scalable manner.
SN - 1471-2105
ER -