Classification and specific primer design for accurate detection of SARS-CoV-2 using deep learning

Abstract

In this paper, deep learning is coupled with explainable artificial intelligence techniques for the discovery of representative genomic sequences in SARS-CoV-2. A convolutional neural network classifier is first trained on 553 sequences from the National Genomics Data Center repository, separating the genome of different virus strains from the Coronavirus family with 98.73% accuracy. The network’s behavior is then analyzed, to discover sequences used by the model to identify SARS-CoV-2, ultimately uncovering sequences exclusive to it. The discovered sequences are validated on samples from the National Center for Biotechnology Information and Global Initiative on Sharing All Influenza Data repositories, and are proven to be able to separate SARS-CoV-2 from different virus strains with near-perfect accuracy. Next, one of the sequences is selected to generate a primer set, and tested against other state-of-the-art primer sets, obtaining competitive results. Finally, the primer is synthesized and tested on patient samples (n = 6 previously tested positive), delivering a sensitivity similar to routine diagnostic methods, and 100% specificity. The proposed methodology has a substantial added value over existing methods, as it is able to both automatically identify promising primer sets for a virus from a limited amount of data, and deliver effective results in a minimal amount of time. Considering the possibility of future pandemics, these characteristics are invaluable to promptly create specific detection methods for diagnostics.

Fast Evaluation of Viral Emerging Risks (FEVER): A computational tool for biosurveillance, diagnostics, and mutation typing of emerging viral pathogens

“…Here, we used an RT-PCR format to evaluate the performance of the FEVER assays as compared to the U.S. CDC assays. The success of RT-PCR as a diagnostic approach relies heavily on the design quality of the primers and probes (57-59) as well as factors such as probe format, the ability to incorporate degenerate nucleotides, the number of target genomes that can be included in the design process to account for genetic diversity, and the ability to avoid cross-reactivity with near-neighbor genomes (60). In the case of SARS-CoV-2 diagnostics, it is also important that assays are able to differentiate infections caused by other common coronaviruses, in which case a false positive could result in unnecessary and costly mitigation procedures (61).…”

The Role of Artificial Intelligence in Fighting the COVID-19 Pandemic

“…The Immunitrack ApS authors, Prachar et al (2020), a company that provides a immunogenicity assessments during drug development, used the neural network method, NetMHC, to predict which peptides will bind, and so identify epitopes for SARS-CoV-2 vaccine. Finally, as for analyzing COVID-19 gene sequences, Lopez-Rincon et al (2021) propose a CNN to classify 553 genome sequences with promising accuracy results.…”

SARS-CoV-2 Tests: Bridging the Gap between Laboratory Sensors and Clinical Applications

“… 63 Once mutated variants are identified, RT-PCR tests can be rapidly adapted to include primers sequences to detect the new strain. 64 , 65 However, the likelihood of false negatives would be expected to increase with increasing number of mutation sites in techniques targeting fewer genomic regions.…”