Application of Artificial intelligence in Covid-19 genome analysis: A review
Application of Artificial intelligence in Covid-19 genome analysis: A review
Zohreh Javanmard,1Marzieh Esmaeili,2,*
1. Department of Health Information Technology, Ferdows School of Paramedical and Health, Birjand University of Medical Sciences, Birjand, Iran 2. Department of Health Information Management, School of Allied Medical Sciences, Tehran University of Medical Sciences, Tehran, Iran
Introduction: Until now, many studies were done to more recognition of coronavirus (Covid-19) and its characteristics. Analysis of the Covid-19 genome is one of the noticeable subjects in this scope. Evaluation of the Covid-19 genome sequencing can provide valuable information about the origin of the virus, its behavior, mutations, and effective therapeutic methods. The use of Artificial Intelligence (AI) techniques is an important method to analyze the Covid-19 genome and extract its features. Deep learning, machine learning, digital signal processing, and deep neural network, are some examples of AI techniques that can be used for the investigation of Covid-19 genome. Therefore, the present study aims to investigate the AI techniques used in Covid-19 genome analysis.
Methods: In this review, a comprehensive search was done in August 2021 through PubMed with the keywords of “Covid-19”, “Genome” and “Artificial Intelligence” alongside their synonyms in the titles and abstracts of the papers published from 2019 to 2021. The original studies which were related to Covid-19 genome analysis by AI techniques were selected; and non-English papers were left out.
Results: The search resulted in 644 papers, among which 21 papers addressing Covid-19 genome analysis by AI techniques were selected based on their titles and abstracts. About 57% (12 of 21) of the studies, used AI techniques to extract hidden patterns, frequent patterns and characteristics of Covid-19 genome sequences. 71% (15 of 21) of the studies, using AI methods and genome information, made predictions about virus mutations, outcomes severity, localization/residency, reservoir host, host proteins that bind to the Covid-19 RNA genome, subsequent structures of virus proteins, genomic similarities of Covid-19 and other viruses, and subsequent sequences of the virus genome. 31% (8 of 21) of the studies, analyzed and identified virus mutations, based on genome sequence information and AI techniques. In most studies, NCBI GenBank (11 of 21) and GISAID Database (6 of 21) have been used to gather genome sequences. Long Short Term Memory (LSTM), random forest classifiers and decision trees were the most widely used AI techniques in the reviewed studies.
Conclusion: The application of artificial intelligence in discovering the genetic features of Covid-19 is very wide, applications like classification and discrimination of Covid-19 disease from similar diseases, extract hidden patterns, investigate the origin of the virus, and predicting the subsequent Covid-19 structures. Therefore, the use of AI alongside genetic is so valuable to the management and control of coronavirus pandemic, as well as the development of vaccines and therapeutic methods.