etection of Staphylococcus aureus in milk samples using impedance spectroscopy and data processing with information visualization techniques and multidimensional calibration space

Author(s)

Soares, Juliana Coatrini ^{[1, 2]} ; Soares, Andrey Coatrini ^[2] ; Popolin-Neto, Mario ^{[3, 4]} ; Paulovich, Fernando Vieira ^{[3, 5]} ; Oliveira Jr, Osvaldo N. ; Caparelli Mattoso, Luiz Henrique ^[2]

Affiliation

^[1] Univ Sao Paulo, Sao Carlos Inst Phys IFSC, BR-13566590 Sao Carlos - Brazil

^[2] Nanotechnol Natl Lab Agr LNNA, Embrapa Instrumentacao, Sao Carlos - Brazil

^[3] Univ Sao Paulo, Inst Math & Comp Sci ICMC, BR-13566590 Sao Carlos - Brazil

^[4] Fed Inst Sao Paulo IFSP, BR-14804296 Araraquara, SP - Brazil

^[5] Dalhousie Univ DAL, Fac Comp Sci FCS, Halifax, NS B3H 4R2 - Canada

• Document type: Journal article
• Source: SENSORS AND ACTUATORS REPORTS; v. 4, NOV 2022.
• Web of Science Citations: 0

Abstract

Early diagnosis of cattle diseases such as mastitis caused by Staphylococcus aureus (S. aureus) can be made effective if on-site detection methods with portable instruments are available. In this work, we fabricated immunosensors based on a layer-by-layer (LbL) film of chitosan and carbon nanotubes coated with a layer of antibodies to detect S. aureus. Using electrical and electrochemical impedance spectroscopies, detection was possible in buffer solutions and in milk with limits of detection which could be as low as 2.6 CFU/mL for milk, sufficient to detect mastitis at early stages. This high sensitivity is ascribed to the specific interactions involving the antibodies, as demonstrated with polarization-modulated infrared reflection absorption spectroscopy (PMIRRAS). The selectivity of the immunosensor was verified by distinguishing S. aureus-containing samples from possible interferents found in milk, for which the interactive document mapping (IDMAP) was employed. Because the interferents affected the spectra, in spite of this distinguishability, we treated the data with a machine learning technique with decision tree models. A multidimensional calibration space was then obtained with rules that permit interpretability and predictability in detecting S. aureus in matrices with high variability as in milk. (AU)

• FAPESP's process: 18/22214-6 - Towards a convergence of technologies: from sensing and biosensing to information visualization and machine learning for data analysis in clinical diagnosis
• Grantee: Osvaldo Novais de Oliveira Junior
• Support Opportunities: Research Projects - Thematic Grants
• FAPESP's process: 18/18953-8 - Nanostructured films applied in microfluidic biosensors to mastitis detection
• Grantee: Andrey Coatrini Soares
• Support Opportunities: Scholarships in Brazil - Post-Doctoral