Modeling SARS-CoV-2 nucleotide mutations as a stochastic process.

Area of Science:

• Genomics and Bioinformatics
• Virology
• Computational Biology

Background:

• The SARS-CoV-2 genome is subject to mutations, influencing its transmissibility and virulence.
• Understanding mutation patterns is crucial for tracking viral evolution and developing effective countermeasures.
• Previous models have not fully captured the spatio-temporal dynamics of SARS-CoV-2 mutations.

Purpose of the Study:

• To analyze SARS-CoV-2 genome sequence mutations using stochastic modeling in both time-series and spatial domains.
• To investigate mutation rate asymmetries and inter-occurrence distances across key variants of concern.
• To provide insights into the biological mechanisms of viral mutation and enhance mutation prediction models.

Main Methods:

• Modeling nucleotide mutations as a stochastic process, incorporating a Markov Chain embedded Poisson random process for mutation rate matrices.
• Developing a spatial gene sequence model to delineate the distribution of mutation inter-occurrence distances.
• Focusing experiments on five key SARS-CoV-2 variants of concern.

Main Results:

• Distinct asymmetries in mutation rates and nucleotide propensities were observed across different strains, with a mean mutation rate of approximately 2 per month.
• Novel biological insights were gained regarding the characteristic distribution of mutation inter-occurrence distances, showing patterns similar to other diseases.
• The study identified specific spatio-temporal dynamics governing SARS-CoV-2 mutations.

Conclusions:

• The findings offer significant insights into the underlying biological mechanisms of SARS-CoV-2 mutations.
• This research advances the accuracy of existing mutation prediction models.
• The study suggests potential for applying similar spatial random process models to characterize mutations in other virus families.