SERS-based lateral flow assay combined with machine learning for highly sensitive quantitative analysis of Escherichia coli O157:H7

In the present study, surface-enhanced Raman scattering–based lateral flow assay (SERS-LFA) strips were applied to promptly and sensitively detect Escherichia coli O157:H7 ( E . coli O157:H7) to ensure food safety. The SERS nanotags were prepared by connecting peculiar monoclonal antibody (McAb) against E . coli O157:H7 directly onto the surfaces of gold-silver core-shell nanostructures loaded with two-layer Raman reporter molecules of 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB). The Raman signal intensity at 1335 cm −1 on the test line (T line) of SERS-LFA strips was detected in the wide range of 10 1 –10 9 colony-forming units/mL (CFU/mL), and regression models based on machine learning were combined to accurately and quantitatively analyze E . coli O157:H7. The limit of detection (LOD) of the extreme gradient boosting regression (XGBR) based on the Raman signal intensity of DTNB was 6.94 × 10 1 CFU/mL for E . coli O157:H7, which was approximately four orders of magnitude lower than that of visual limits. In addition, although E . coli O157:H7 was spiked into the food matrices including milk and beef at an ultra-low dose of 10 CFU/mL, the SERS-LFA combined with XGBR was able to successfully explore E . coli O157:H7 from the mixture that was incubated for only 2 h, in which the recoveries were mainly distributed between 86.41 and 128.25%. In summary, these results demonstrated that the SERS-LFA had a significant potential as a powerful tool for the point-of-care testing (POCT) of E . coli O157:H7 in the early food contamination stage.