Structured illumination microscopy combined with machine learning enables the high throughput analysis and...

Area of Science:

• Biophysics
• Microscopy
• Machine Learning

Background:

• Optical super-resolution microscopy provides high spatial resolution for studying supramolecular structures like viruses.
• Investigating virus morphology is crucial for understanding their function, particularly in biopharmaceutical applications.

Purpose of the Study:

• To develop and demonstrate a new methodology combining Structured Illumination Microscopy (SIM) with machine learning for high-resolution imaging and classification of virus populations.
• To link virus morphology data with functional performance for improved biopharmaceutical production.
• To establish a rapid, high-throughput quality assessment tool for viruses.

Main Methods:

• Structured Illumination Microscopy (SIM) was employed for high-resolution imaging.
• Machine learning algorithms were integrated with SIM for automated image analysis and virus classification.
• The methodology was tested on Newcastle Disease Virus (oncolytic virotherapy) and Influenza virus (vaccine development).

Main Results:

• The combined SIM and machine learning approach successfully imaged and classified large populations of viruses with high resolution.
• The method provided morphological information potentially linkable to functional performance.
• The approach demonstrated efficacy on non-purified samples directly from production lines.

Conclusions:

• This novel methodology offers a powerful tool for rapid, high-throughput assessment of biopharmaceutical virus quality.
• It obviates the need for complex and time-consuming traditional batch testing methods.
• The technique has significant implications for oncolytic virotherapy and vaccine development, streamlining production and quality control.