Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Neuromorphic Spiking Neural Network Based Classification of COVID-19 Spike Sequences

Taslim Murad, Prakash Chourasia, Sarwan Ali, Imdad Ullah Khan, Murray Patterson

TL;DR

This work addresses the challenge of classifying SARS-CoV-2 spike protein sequences amid rapid viral mutations by proposing an alignment-free pipeline that uses a neuromorphic spiking neural network (SNN). It converts spike sequences into fixed-length numerical representations and processes them with a two-layer SNN that handles temporal dynamics, trained end-to-end with cross-entropy loss. Evaluated on the Spike7k dataset against multiple baselines, the SNN achieves higher accuracy and competitive performance across key metrics, demonstrating the effectiveness and efficiency of neuromorphic approaches for complex biological sequences. The study highlights the spike region as a rich target for analysis and suggests that neuromorphic computing can offer practical gains in sequence classification tasks with varying lengths and imbalanced classes.

Abstract

The availability of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) virus data post-COVID has reached exponentially to an enormous magnitude, opening research doors to analyze its behavior. Various studies are conducted by researchers to gain a deeper understanding of the virus, like genomic surveillance, etc, so that efficient prevention mechanisms can be developed. However, the unstable nature of the virus (rapid mutations, multiple hosts, etc) creates challenges in designing analytical systems for it. Therefore, we propose a neural network-based (NN) mechanism to perform an efficient analysis of the SARS-CoV-2 data, as NN portrays generalized behavior upon training. Moreover, rather than using the full-length genome of the virus, we apply our method to its spike region, as this region is known to have predominant mutations and is used to attach to the host cell membrane. In this paper, we introduce a pipeline that first converts the spike protein sequences into a fixed-length numerical representation and then uses Neuromorphic Spiking Neural Network to classify those sequences. We compare the performance of our method with various baselines using real-world SARS-CoV-2 spike sequence data and show that our method is able to achieve higher predictive accuracy compared to the recent baselines.

Neuromorphic Spiking Neural Network Based Classification of COVID-19 Spike Sequences

TL;DR

This work addresses the challenge of classifying SARS-CoV-2 spike protein sequences amid rapid viral mutations by proposing an alignment-free pipeline that uses a neuromorphic spiking neural network (SNN). It converts spike sequences into fixed-length numerical representations and processes them with a two-layer SNN that handles temporal dynamics, trained end-to-end with cross-entropy loss. Evaluated on the Spike7k dataset against multiple baselines, the SNN achieves higher accuracy and competitive performance across key metrics, demonstrating the effectiveness and efficiency of neuromorphic approaches for complex biological sequences. The study highlights the spike region as a rich target for analysis and suggests that neuromorphic computing can offer practical gains in sequence classification tasks with varying lengths and imbalanced classes.

Abstract

The availability of SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) virus data post-COVID has reached exponentially to an enormous magnitude, opening research doors to analyze its behavior. Various studies are conducted by researchers to gain a deeper understanding of the virus, like genomic surveillance, etc, so that efficient prevention mechanisms can be developed. However, the unstable nature of the virus (rapid mutations, multiple hosts, etc) creates challenges in designing analytical systems for it. Therefore, we propose a neural network-based (NN) mechanism to perform an efficient analysis of the SARS-CoV-2 data, as NN portrays generalized behavior upon training. Moreover, rather than using the full-length genome of the virus, we apply our method to its spike region, as this region is known to have predominant mutations and is used to attach to the host cell membrane. In this paper, we introduce a pipeline that first converts the spike protein sequences into a fixed-length numerical representation and then uses Neuromorphic Spiking Neural Network to classify those sequences. We compare the performance of our method with various baselines using real-world SARS-CoV-2 spike sequence data and show that our method is able to achieve higher predictive accuracy compared to the recent baselines.
Paper Structure (15 sections, 2 equations, 2 figures, 2 tables)

This paper contains 15 sections, 2 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Proposed Approach
  4. Experimental Evaluation
  5. Dataset Statistics
  6. Evaluation Metrics and Classifiers
  7. Baseline Methods
  8. Autoencoder + Neural Tangent Kernel nguyen2021benefits
  9. Image Transformation
  10. Recurrent Plot (RP)
  11. Gramian Angular Field (GAF)
  12. Markov Transition Field (MTF)
  13. Signal Transformation
  14. Results and Discussion
  15. Conclusion

Figures (2)

  • Figure 1: The genome of SARS-CoV-2 has a length of 30kb, and it consists of non-structural proteins (ORFs 1ab) and structural proteins (E, M, N, S). The S region is important because of its ability to attach to the host cell membrane, and also hold advantageous mutations.
  • Figure 2: The images from ($a-c$) are against a spike sequence corresponding to Lineage AY.12, while ($d-f$) show images of a sequence from B.1.526 Lineage. They are created using RP, GAF, and MTF approaches.