GenEx Detects SARS-CoV-2 Variants Using Codon Co-occurrence Graphs.

Arefin Amin, Labiba Faiza Karim, M. Monir Uddin· August 20, 2026 View original

Key takeaways

  • GenEx uses codon co-occurrence graphs to capture complex genomic interdependencies for variant detection.
  • The method outperforms traditional linear sequence analysis in identifying SARS-CoV-2 variants.
  • Spectral graph feature extraction with squared singular values enhances classification accuracy.
  • This paradigm has potential applications beyond viral genomics.

Who benefits

HealthcarePharmaceuticalsBiotechnologyPublic HealthResearch & Development

Summary

GenEx is a new pipeline that converts viral gene sequences into codon co-occurrence graphs, extracting over 25 graph features to detect SARS-CoV-2 variants. This graph-based approach, inspired by computational linguistics, outperforms classical bioinformatics methods by capturing complex contextual interdependencies in genomic data.

This paper introduces GenEx, an innovative pipeline designed for the detection of SARS-CoV-2 variants, such as Beta, Gamma, Delta, and Omicron. Diverging from traditional bioinformatics methods that treat gene sequences as linear strings, GenEx transforms raw gene sequences into codon co-occurrence graphs. This approach allows for the extraction of over 25 distinct graph features, capturing the intricate contextual interdependencies within the genomic data, a capability often missed by sequence alignment or phylogenetic analysis. The core of GenEx involves two key techniques: Multi-Scale Codon Co-occurrence Graph (MSCG) and Linear-time Adjacency PMI Codon Graph (LAPCG). These algorithms enable the interpretation of codon sequences as structured symbolic vocabularies, drawing inspiration from computational linguistics. A significant contribution is the use of Singular Value Decomposition (SVD) for spectral graph feature extraction, specifically utilizing the squared singular value (σ²) to enhance the separation between dominant and subdominant spectral components, thereby improving inter-class separability for classification. The efficacy of GenEx was demonstrated by training 23 benchmarked machine learning models, which achieved remarkable accuracy in detecting all tested SARS-CoV-2 variants.

Why it matters

This novel graph-based approach offers a more accurate and nuanced method for viral variant detection, potentially accelerating public health responses and drug development.

How to implement this in your domain

  1. 1Evaluate the GenEx pipeline for rapid and accurate detection of new viral variants in genomic surveillance programs.
  2. 2Collaborate with bioinformatics teams to integrate graph-based feature extraction into existing genomic analysis workflows.
  3. 3Explore applying the codon co-occurrence graph paradigm to other areas of genomic research, such as bacterial strain identification or cancer genomics.
  4. 4Investigate the use of SVD with squared singular values for feature extraction in other high-dimensional biological datasets.

Original post by Arefin Amin, Labiba Faiza Karim, M. Monir Uddin

"arXiv:2608.18238v1 Announce Type: new Abstract: Genomic analysis on viruses such as SARS-CoV-2 variants: Beta, Gamma, Delta, and Omicron is heavily dominated by classical bioinformatics methods, including Sequence Alignment, Phylogenetic Analysis, and Mutation Frequency Statistic…"

View on X

Originally posted by Arefin Amin, Labiba Faiza Karim, M. Monir Uddin on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses