A rapid method for determination of plasmid types and transformations based on combining the Fourier-transform infrared spectral data with machine learning

Ruili Li, Wenlong Liang, Fengfeng Li, Fangxiao Li, Lingling Tao, Shouning Yang, Chunlei Wang, Huayan Yang,
A rapid method for determination of plasmid types and transformations based on combining the Fourier-transform infrared spectral data with machine learning,
Microchemical Journal,
Volume 218,
2025,
115287,
ISSN 0026-265X,
https://doi.org/10.1016/j.microc.2025.115287.
(https://www.sciencedirect.com/science/article/pii/S0026265X25026359)
Abstract: Plasmid transformation, a fundamental technique widely employed in biological research, faces significant bottlenecks in rapid verification. Fourier-transform infrared (FTIR) spectroscopy provides a non-destructive analytical method, enabling sensitive detection of functional group-specific absorption peaks and high-throughput spectral acquisition from minimally processed biological samples. Machine learning algorithms can enhance spectral data analysis to address challenges in biological sample identification. Leveraging the high-throughput capability of FTIR, this study developed a rapid method combining FTIR spectroscopy and machine learning to classify muscarinic acetylcholine receptor (M receptor family) plasmid subtypes and their harboring Escherichia coli (E. coli) cells. Final results confirm that analysis using full-spectrum data (400–4000 cm−1) delivers optimal performance, with all core classification algorithms consistently exceeding 80 % accuracy for both plasmid subtypes and plasmid-transformed E. coli cells. This approach presents a time- and cost-efficient alternative to traditional biochemical assays for verifying plasmid transformation success, requiring less technical expertise. It establishes a promising pathway for the rapid classification of diverse biological samples.
Keywords: FTIR; Machine learning; Plasmid classification; Plasmid transformation