Improved Antibiotic Detection in Raw Milk Using Machine Learning Tools over the Absorption Spectra of a Problem-Specific Nanobiosensor

Abstract

In this article we present the development of a biosensor system that integrates nanotechnology, optomechanics and a spectral detection algorithm for sensitive quantification of antibiotic residues in raw milk of cow. Firstly, nanobiosensors were designed and synthesized by chemically bonding gold nanoparticles (AuNPs) with aptamer bioreceptors highly selective for four widely used antibiotics in the field of veterinary medicine, namely, Kanamycin, Ampicillin, Oxytetracycline and Sulfadimethoxine. When molecules of the antibiotics are present in the milk sample, the interaction with the aptamers induces random AuNP aggregation. This phenomenon modifies the initial absorption spectrum of the milk sample without antibiotics, producing spectral features that indicate both the presence of antibiotics and, to some extent, its concentration. Secondly, we designed and constructed an electro-opto-mechanic device that performs automatic high-resolution spectral data acquisition in a wavelength range of 400 to 800 nm. Thirdly, the acquired spectra were processed by a machine-learning algorithm that is embedded into the acquisition hardware to determine the presence and concentration ranges of the antibiotics. Our approach outperformed state-of-the-art standardized techniques (based on the 520/620 nm ratio) for antibiotic detection, both in speed and in sensitivity.

A Simple Yet Effective Preanalytical Strategy Enabling the Application of Aptamer-Conjugated Gold Nanoparticles for the Colorimetric Detection of Antibiotic Residues in Raw Milk

“…The shift of the SPR peak (from 520 nm) associated with the aggregation of AuNPs spans a wide spectral region (until 800 nm), despite most of the works in the field of NAS following the shift between 520 nm and 620 or 650 nm [ 50 ]. Accordingly, our group previously developed an electro-opto-mechanic device for high-resolution AuNP spectral data in a wavelength range from 400 to 800 nm, which was proven to improve the analytical performance of NAS with the aid of machine learning tools [ 51 ]. With this in mind, AuNP aggregation data were analyzed spectrophotometrically between 400 and 750 nm, using two absorption ratios (A 520 /A 620 and A 520 /A 720 ) to select reading parameters that maximize the colorimetric signal associated with NAS detection and aggregation.…”

Toxicity prediction based on artificial intelligence: A multidisciplinary overview

“…However, the use of these techniques entails offline limitations, time, and destruction. To decrease the analysis times, emerging non‐destructive methodologies as the spectroscopic technologies include hyperspectral imaging 139 , fluorescence spectroscopy 140 , near‐infrared (IR) spectroscopy, Fourier transform IR (FTIR) and Raman spectroscopies, biosensors 141 , electronic‐nose, 142 and electronic‐tongue 143 have been introduced as promising techniques for the analysis of several of these compounds. All these techniques are currently used in combination with different AI algorithms, for example, ML, DL, and different classes including those within these architectures, such as ANNs, CNN, DNN, and so on, for toxicity issues, specifically predicting the concentrations of different harmful compounds in various food matrices (see Table 2).…”

“…Aflatoxins, a big issue from the point of view of food safety, were identified and quantified in almonds (total Aflatoxins B, AFB1 + AFB2) using a simple portable fluorescence spectroscopy complemented with AI 140 . Remarkably, a portable biosensor system integrating SVM algorithms recently provided good results with an on‐site and sensitive detection of a number of antibiotic residues in cow milk 141 . Also, a rapid on‐line detection method was demonstrated for beverage preservatives benzoic acid, potassium sorbate, sodium dehydrogenate, and propyl p ‐hydroxybenzoate, based on X‐ray absorption spectroscopies (XAS), complemented with DNN and SVM classifiers for data classification and prediction, achieving fast on‐line detection above the standard levels, 150 while ML methods perfectly complemented terahertz (THz) spectroscopy techniques for rapidly measuring the benzoic acid additive in wheat flour, with a good correlation coefficient and low root mean square error of prediction 151 .…”

“…140 Remarkably, a portable biosensor system integrating SVM algorithms recently provided good results with an on-site and sensitive detection of a number of antibiotic residues in cow milk. 141 Also, a rapid on-line detection method was demonstrated for beverage preservatives benzoic acid, potassium sorbate, sodium dehydrogenate, and propyl p-hydroxybenzoate, based on X-ray absorption spectroscopies (XAS), complemented with DNN and SVM classifiers for data classification and prediction, achieving fast on-line detection above the standard levels, 150 while ML methods perfectly complemented terahertz (THz) spectroscopy techniques for rapidly measuring the benzoic acid additive in wheat flour, with a good correlation coefficient and low root mean square error of prediction. 151 E-tongue and e-nose have been reported as novel low-cost and simple optical electronic systems when coupled with ML approaches, with good correlation coefficients and low mean squares errors for the simultaneous and quantitative prediction of Pb, Cd, and Hg residues in fish (Carassius carassius) samples, 143 and fast and cost-effective detection of food additives benzoic acid and chitosan in fruit juices.…”

A Novel Preanalytical Strategy Enabling Application of a Colorimetric Nanoaptasensor for On-Site Detection of AFB1 in Cattle Feed

“…AuNP-based NASs enable colorimetric detection by following a red-to-purple blue shift of the absorption spectrum during the aggregation of the nanoparticles, providing a versatile, sensitive, selective, and easy-to-apply platform. We recently reported NAS systems for the colorimetric detection of antibiotic residues from raw milk samples, addressing both the preanalytical problem [ 39 ] and the application of a machine-learning algorithm to improve the sensitivity of the analytical process [ 43 ]. Aptasensors for AFB1 detection have been successfully reported in rice, peanuts, and wheat flour upon extraction using methanol/water in 80/20 and 60/40 mixtures [ 19 , 44 ].…”