# Milk NIR spectroscopy and Aquaphotomics novel diagnostic approach to Paratuberculosis in dairy cattle

## TL;DR

This paper explores using milk near-infrared spectroscopy and Aquaphotomics to detect Paratuberculosis in dairy cattle, offering a non-destructive and accurate alternative to traditional methods.

## Contribution

The study introduces a novel non-destructive diagnostic approach using NIRS and Aquaphotomics for early detection of Paratuberculosis in dairy cattle.

## Key findings

- NIRS and Aquaphotomics achieved 100% accuracy in internal validation using quadratic discriminant analysis and support vector machine models.
- External validation showed 100% sensitivity when using blood plasma ELISA as a reference test.
- Milk NIR spectra can indirectly analyze antibody levels through changes in water spectral patterns.

## Abstract

Mycobacterium Avium subspecies Paratuberculosis (MAP) causes Johne’s disease or Paratuberculosis, a chronic, progressive intestinal disease in ruminants. The incidence and prevalence of Johne’s disease are higher in dairy cattle herds because of intensive breeding and high production. Developing non-destructive diagnostic methods for early detection of this disease by simple sampling is paramount for breeding, economic, and health programs. Conventional methods are almost entirely destructive, have low accuracy, and are time-consuming. Near- infrared spectroscopy (NIRS) and Aquaphotomics can detect changes in biofluids and thus have the potential to diagnose the disease. This study aimed to investigate the diagnostic ability of NIRS and Aquaphotomics for Paratuberculosis in dairy cattle by milk sample. Milk samples from dairy cattle were collected in the NIR range (1300–1600 nm) 60 days before and 100–200 days after calving in two groups, positive and negative, using the three same consecutive ELISA test results of blood plasma and milk, as a reference test. The NIRS and Aquaphotomics methods in quadratic discriminant analysis (QDA) and support vector machine (SVM) models achieved high accuracy in detecting negative and positive groups. In internal validation, SVM and QDA models in 12 water absorbance bands had 100% accuracy. In external validation, milk samples with blood plasma ELISA reference test achieved 100% sensitivity, which is more accurate than milk ELISA as a reference test. The current study found that monitoring milk with NIR spectra provides an opportunity to analyze antibody levels indirectly via changes in water spectral patterns caused by complex physiological changes, such as the amount of antibodies related to Paratuberculosis by aquagram.