Automatic 3D cluster modelling of COVID-19 through voxel-based redistribution

Area of Science:

• Computational Biology
• Biophysics
• Virology

Background:

• Studying virus dynamics computationally offers a non-contact research environment.
• Accurate virus cluster modeling is essential for understanding group virus properties.
• Existing methods struggle with the morphological complexity of viruses, limiting smooth function algorithms.

Purpose of the Study:

• To develop an automatic, voxel-based approach for generating virus clusters from 3D data.
• To enable accurate modeling of complex virus morphologies, specifically for SARS-CoV-2 (COVID-19).
• To demonstrate the necessity of including structural details, like spike proteins, in cluster models for simulations.

Main Methods:

• Proposed a voxel-based redistribution approach for automatic virus cluster generation.
• Utilized Representative Elementary Volume (REV) analysis for statistical robustness.
• Performed coordination number analysis and surface density measurements for comparison with spherical models.

Main Results:

• The voxel-based approach successfully generated virus clusters reflecting COVID-19 morphology.
• REV analysis confirmed statistical reliability of the digital samples.
• Virtual permeation simulations highlighted significant differences between COVID-19 and spherical clusters, emphasizing the need for structural accuracy.

Conclusions:

• The proposed voxel-based method effectively models complex virus structures, overcoming limitations of traditional algorithms.
• Accurate cluster modeling, including spike protein details, is critical for reliable fluid dynamics and permeation simulations.
• This approach enhances computational studies of virus dynamics and interactions.