Related Experiment Video
Updated: Apr 6, 2026
Multicolor Fluorescence Detection for Droplet Microfluidics Using Optical Fibers
Published on: May 5, 2016
Machine learning-assisted Zr-MOFFe@SECHA fluorescence and colorimetric dual-mode intelligent platform for sensitive
Pengying Liang1, Wenhong Zhou1, Jun Li1
1Department of Medicine, the Sixth Affiliated Hospital of South China University of Technology (Nanhai District People's Hospital of Foshan), Foshan, 528200, PR China.
Background:
Lipopolysaccharide (LPS) is a crucial diagnosis biomarker for sepsis. Accurate and sensitive determination of LPS in real samples (human serum and vancomycin) assumes paramount importance in effectively preventing sepsis and promoting public health. Regrettably, traditional detection methods often suffer from a single signal output, which is insufficient for accurate and reliable analysis of LPS in real samples. Therefore, it remains a challenge to develop an accurate and sensitive LPS analysis strategy to facilitate early diagnosis and drug safety.
Results:
The integration of self-feedback exponential catalytic hairpin assembly (SECHA) with Zr-MOFFe was proposed (denoted as Zr-MOFFe@SECHA) to achieve fluorescence and colorimetric dual-mode for the accurate and sensitive analysis of LPS in real samples. The optimally synthesized Zr-MOFFe exhibited satisfactory peroxidase activity and enhanced fluorescence quenching efficiency. The SECHA system consisted of two hairpin probes (H1 and H2) and two double-stranded DNA probes (ST and HP). The rational design endowed the system with remarkable exponential amplification efficiency. Both modes exhibited broad linear ranges (1 fg mL-1 to 1 ng mL-1 for fluorescence and 5 pg mL-1 to 5 ng mL-1 for colorimetry) and lower detection limits (0.254 fg mL-1 and 0.143 pg mL-1, respectively). Finally, with the help of machine learning, the backpropagation neural network (BPNN) model could enable intelligent, accurate, and sensitive determination of LPS concentration based on fluorescence spectra and colorimetric images.
Significance:
The experimental results indicate that the constructed strategy is a sensitive and reliable platform for monitoring LPS. It exhibits high selectivity and reproducibility toward different analytes. Also, the applicability in real samples has been verified. We believe that the proposed machine learning-assisted Zr-MOFFe@SECHA method will carry significant implications for LPS-related disease diagnosis and drug safety.
