Hamlet-Pattern-Based Automated COVID-19 and Influenza Detection Model Using Protein Sequences.

Area of Science:

• Bioinformatics
• Computational Biology
• Infectious Disease Diagnostics

Background:

• SARS-CoV-2 and Influenza-A share overlapping symptoms, complicating clinical diagnosis.
• The ongoing COVID-19 pandemic necessitates reliable methods to differentiate between these viral infections.
• Seasonal influenza continues to be a public health concern alongside SARS-CoV-2.

Purpose of the Study:

• To develop and evaluate a novel text-based classification model for discriminating between SARS-CoV-2 and Influenza-A.
• To leverage viral protein sequences for accurate and efficient differential diagnosis.
• To create a computer-aided diagnostic tool for screening these respiratory viruses.

Main Methods:

• A dataset of 16,901 SARS-CoV-2 and 19,523 Influenza-A protein sequences was compiled from the NCBI database.
• A novel feature extraction method, HamletPat, was employed to generate binary patterns from protein sequences.
• Support Vector Machine (SVM) with a Gaussian kernel, utilizing 340 selected features, performed the classification.

Main Results:

• The classification model achieved high accuracy rates: 99.92% with hold-out validation and 99.87% with five-fold cross-validation.
• The HamletPat feature extraction method proved effective in capturing discriminative information from protein sequences.
• A lightweight, handcrafted feature set enabled efficient and accurate classification.

Conclusions:

• The developed HamletPat-based classification model demonstrates excellent performance in distinguishing SARS-CoV-2 from Influenza-A.
• This model offers a valuable and efficient tool for screening viral protein sequences for diagnostic purposes.
• Computer-aided diagnosis using protein sequence analysis can significantly aid in managing co-circulating respiratory viral infections.