Artificial intelligence enhanced electrochemical immunoassay for staphylococcal enterotoxin B.

Journal: Scientific reports
Published Date:

Abstract

Staphylococcal enterotoxin B (SEB) holds critical importance in disease diagnosis, food safety, and public health due to its high toxicity and potent pathogenicity. Traditional immunoassay methods for detecting SEB often exhibit insufficient accuracy and robustness. This study leverages machine learning technology to integrate the quantitative measurement advantages of electrochemical methods with the strong specificity of immunoassays, achieving high-precision coupled electrochemical immunodetection of SEB. Firstly, an electrochemical immunosensing system was developed to capture the target analyte SEB by immobilizing specific antibodies on the electrode surface. Cyclic voltammetry (CV) was utilized to accurately characterize the immune response process. Secondly, feature selection methodologies within machine learning are utilized to identify eight key parameters from CV curves that are highly related to SEB concentration. This enhancement significantly improves both the accuracy and interpretability of SEB measurement data. Lastly, a multivariate linear regression algorithm is employed to effectively train and fit the extracted feature data. This approach successfully mitigates noise introduced by variations in electrode batches, experimental conditions, and operational techniques-thereby enabling robust quantitative measurements of SEB concentration with high precision. The entire detection process requires only 20 μL sample and is accomplished in just two minutes. This method can detect antigen concentrations at both ng/mL and μg/mL levels, with a detection limit of 1 ng/mL. The [Formula: see text] score for predicting SEB antigen concentration is approximately 0.999, accompanied by a mean absolute percentage error (MAPE) of 6.09% This approach achieves high precision, robustness, and specificity in SEB detection, offering extensive detection range, rapid response time, and cost-effectiveness, presenting new opportunities for identifying various pathogenic toxins.

Authors

  • Yuliang Zhao
    Department of Control Engineering, Northeastern University, Qinhuangdao, Hebei, 066001, PR China.
  • Tingting Sun
    Center for Systems Biology, Department of Bioinformatics, School of Biology and Basic Medical Sciences, Soochow University, 215123, Suzhou, China.
  • Huawei Zhang
    Department of Ultrasound, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan, China.
  • Chao Lian
    School of Control Engineering, Northeastern University at Qinhuangdao, Qinhuangdao, 066000, Hebei, China.
  • Zhongpeng Zhao
    Department of Immunology, School of Basic Medical Sciences, Anhui Medical University, No 81 Meishan Road, Shushan District, Hefei 230032, China.
  • Yongqiang Jiang
    State Key Laboratory of Pathogen and Biosecurity, Academy of Military Medical Sciences, Beijing, 100071, China.
  • Huiqi Duan
    State Key Laboratory of Pathogen and Biosecurity, Academy of Military Medical Sciences, Beijing, 100071, China.
  • Yuhao Ren
    State Key Laboratory of Pathogen and Biosecurity, Academy of Military Medical Sciences, Beijing, 100071, China.
  • Xuyang Sun
    Department of Radiology, Tokai University School of Medicine, 143 Shimokasuka, Isehara, 259-1193, Japan.
  • Zhikun Zhan
    School of Control Engineering, Northeastern University at Qinhuangdao, Qinhuangdao, 066000, Hebei, China. zhanzhikun@neuq.edu.cn.
  • Mingyue Qu
    The PLA Rocket Force Characteristic Medical Center, Beijing, 100088, China. qumingyue2008@126.com.
  • Shaolong Chen
    School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen, 518107, China.