Indirect Detection of Swine Influenza Activity in Porcine Blood Using Raman Spectroscopy and Machine Learning

Area of Science:

• Veterinary Medicine
• Biotechnology
• Spectroscopy

Background:

• Swine influenza variants like H1N1 and H1N2 are recurrent threats to animal health.
• Current detection methods may lack speed or require sample manipulation.
• Raman spectroscopy (RS) presents a label-free, non-destructive technique for pathogen detection.

Purpose of the Study:

• To evaluate the efficacy of Raman spectroscopy (RS) for identifying swine influenza virus (SIV) variants in porcine blood serum.
• To assess the performance of RS combined with machine learning algorithms for infection status determination.

Main Methods:

• Blood serum samples from healthy, vaccinated, and SIV-infected swine (H1N1, H1N2) were analyzed using Raman spectroscopy.
• Spectral data were processed using machine learning algorithms, including partial least squares discriminant analysis (PLS-DA) and eXtreme gradient boosting discriminant analysis (XGB-DA).

Main Results:

• The combined RS and machine learning approach achieved high accuracy rates, up to 97.8%, in identifying infection status.
• Specific SIV variants (H1N1, H1N2) were distinguishable within porcine blood serum samples.
• The method demonstrated effectiveness across different swine conditions (healthy, vaccinated, infected).

Conclusions:

• Raman spectroscopy is a powerful and novel tool for rapid, label-free detection of swine influenza variants.
• This technique significantly improves the potential for enhanced surveillance and early identification of animal health threats in swine populations.
• The integration of RS with machine learning offers a promising diagnostic strategy for veterinary applications.