Journal of Forensic Medicine ›› 2025, Vol. 41 ›› Issue (5): 468-476.DOI: 10.12116/j.issn.1004-5619.2025.550803

• Topic on Microbiomics • Previous Articles    

Exploring Microbial Detection of Psychoactive Substances in Wastewater Based on Micobiome Analysis

Hou-ying ZOU1(), Yin-lei LEI2,3(), Ruo-cheng XIA2, Yan SHI2, Cheng-tao LI1,2,4()   

  1. 1.School of Forensic Medicine, Southern Medical University, Guangzhou 510515, China
    2.Shanghai Key Laboratory of Forensic Medicine, Key Laboratory of Forensic Science, Ministry of Justice, Shanghai Forensic Service Platform, Academy of Forensic Science, Shanghai 200063, China
    3.School of Preclinical Medicine of Zunyi Medical University, Zunyi 563000, Guizhou Province, China
    4.Institute of Forensic Science, Fudan University, Shanghai 200032, China
  • Received:2025-08-18 Online:2026-01-27 Published:2025-10-25
  • Contact: Cheng-tao LI

Abstract:

Objective To explore the potential wastewater microbiome analysis for detecting psychoactive substances by using full-length 16S rRNA gene sequencing with liquid chromatography - tandem mass spectrometry (LC-MS/MS). Methods LC-MS/MS was used to qualitatively detect psychoactive substances in 21 wastewater samples suspected to contain such compunds. Based on the results, the samples were categorized into two groups: a positive group (containing psychoactive substances) and a negative group (free of psychoactive substances). Subsequently, bacterial communities in all samples were analyzed using full-length 16S rRNA gene sequencing. This analysis characterized the species composition, α diversity (Shannon and Simpson indices), and β-diversity (PCoA and NMDS). Significantly different operational taxonomic units (OTUs) were screened using linear discriminant analysis effect size (LEfSe), and optimal OTU features were iteratively selected via recursive feature elimination (RFE). A random forest prediction model was built with these two OTU subsets as input features. Results The composition and structure of the bacterial communities showed marked differences between the two groups. The sample diversity in the positive group was higher than that in the negative group. The permutational ultivariate analysis of variance (PERMANOVA) revealed a statistically significant difference in β-diversity between the two groups. Random Forest models achieved a prediction accuracy of 83.3%, with areas under the ROC curve of 0.89 and 0.83, respectively. Conclusion Integrating wastewater bacterial community analysis with chemical analysis techniques may provide a more comprehensive approach for monitering the presence of psychoactive substances.

Key words: forensic medicine, microbiomics, psychoactive substances, machine learning, 16S rRNA

CLC Number: