What question did this study set out to answer?

The aim is to develop KnvResGAT for accurate classification of SARS-CoV-2 lineages using novel representation and neural network techniques.

February 14, 2026Open Access

KnvResGAT: SARS-CoV-2 sequence classification using k-mer natural vector and graph attention networks

Key Points

The aim is to develop KnvResGAT for accurate classification of SARS-CoV-2 lineages using novel representation and neural network techniques.
Combined k-mer Natural Vector representations with a residual multi-head Graph Attention Network.
Constructed a k-nearest-neighbor similarity graph in the feature space of KNV.
Evaluated on a dataset of 182,851 curated SARS-CoV-2 genomes across 103 Pango lineages.
Achieved 0.9729 accuracy and 0.9636 Macro-F1 score.
Outperformed Pangolin accuracy (0.9673) and Macro-F1 (0.9471).
Showed better performance than the deep baseline ResMLP (0.9654 accuracy, 0.9520 Macro-F1).

Abstract

Abstract Objective We propose KnvResGAT for efficient SARS-CoV-2 lineage classification by combining k-mer Natural Vector (KNV) representations with a residual multi-head Graph Attention Network (GAT) on a k-nearest-neighbor (kNN) similarity graph constructed in the KNV feature space. Results On a time-aware per-lineage split of 182,851 curated SARS-CoV-2 genomes spanning 103 Pango lineages, KnvResGAT achieved 0.9729 accuracy and 0.9636 Macro-F1. Under the same split, it outperformed Pangolin (0.9673 accuracy, 0.9471 Macro-F1) and a strong deep baseline ResMLP (0.9654 accuracy, 0.9520 Macro-F1), demonstrating improved generalization for multi-class lineage classification.

Bookmark

View Full Paper

Bookmark

View Full Paper

KnvResGAT: SARS-CoV-2 sequence classification using k-mer natural vector and graph attention networks

Key Points

Abstract

Cite This Study