Abstract
AbstractCOVID-19, currently prevalent worldwide, is caused by a novel coronavirus, SARS-CoV-2. Similar to other RNA viruses, SARS-CoV-2 continues to evolve through random mutations, creating numerous variants, such as Alpha, Beta, and Delta. It is, therefore, necessary to predict the mutations constituting the dominant variant before they are generated. This can be achieved by continuously monitoring the mutation trends and patterns. Hence, in the current study, we sought to design a dominant variant candidate (DVC) selection algorithm. To this end, we obtained COVID-19 sequence data from GISAID and extracted position-nucleotide (POS-NT) frequency ratio data by country and date through data preprocessing. We then defined the dominant dates for each variant in the USA and developed a frequency ratio prediction model for each POS-NT. Based on this model, we applied DVC criteria to develop the selection algorithm, verified for Delta and Omicron. Using Condition 3 as the DVC criterion, 69 and 102 DVC POS-NTs were identified for Delta and Omicron an average of 47 and 82 days before the dominant dates, respectively. Moreover, 13 and 44 Delta- and Omicron-defining POS-NTs were recognized 18 and 25 days before the dominant dates, respectively. We identified all DVC POS-NTs before the dominant dates, including soaring and gently increasing POS-NTs. Considering that we successfully defined all POS-NT mutations for Delta and Omicron, the DVC algorithm may represent a valuable tool for providing early predictions regarding future variants, helping improve global health.Author Summary
Publisher
Cold Spring Harbor Laboratory
Reference12 articles.
1. Rasmussen CE. Gaussian Processes in Machine Learning. In: Bousquet O, von Luxburg U, Rätsch G, editors. Summer School on Machine Learning. Springer; 2004. pp. 63–71.
2. A tutorial on Gaussian process regression: Modelling, exploring, and exploiting functions
3. Jarndal A , Husain S , Zaatar O , Al Gumaei T , Hamadeh A. In: 2020 International Conference on Communications, Computing, Cybersecurity, and Informatics (CCCI); 2020. pp. 1–5.
4. GHI forecasting using Gaussian process regression: Kernel study;IFAC-PapersOnLine,2019
5. Velásquez RMA , Lara J VM . Forecast and evaluation of COVID-19 spreading in USA with reduced-space Gaussian process regression. Chaos Solit Fractals 2020;136: 109924.