Thanatomicrobiome and artificial intelligence: forensic microbiology today

• Autores: Baltazar Ramos, Javier Iván, Cosme García , Lizbeth, Denis Rodriguez, Edmundo
• Formato: artículo
• Fecha de Publicación: 2025
• Institución: Universidad de San Martín de Porres
• Repositorio: Horizonte médico
• Lenguaje: español
• OAI Identifier: oai:horizontemedico.usmp.edu.pe:article/3758
• Enlace del recurso: https://www.horizontemedico.usmp.edu.pe/index.php/horizontemed/article/view/3758
• Nivel de acceso: acceso abierto
• Materia: Microbiología Forense; Cambios Post Mortem; Inteligencia Artificial; Aprendizaje Profundo; Cadáver; Forensic Microbiology; Postmortem Changes; Artificial Intelligence; Deep Learning; Cadaver

Forensic microbiology enables, among other applications, the estimation of the post-mortem interval (PMI), the identification of individuals, and the location of crime scenes through microbiome analysis and the geolocation of biological remains. Artificial intelligence (AI), together with new sequencing techniques, has revolutionized this field, markedly improving the accuracy and speed of forensic analyses. In this study, a systematic review was conducted following PRISMA guidelines. Databases such as PubMed, Scopus, Web of Science, and Google Scholar were searched using keywords related to forensic microbiology, IA, and PMI. Inclusion criteria included studies published in English or Spanish, regardless of the publication date. Exclusion criteria included duplicate studies or those that did not address the thanatomicrobiome analysis using AI tools. After the search and selection process, 20 articles published between 2016 and 2024 were analyzed. The f indings show that some machine learning models, such as Random Forest (RF) and Convolutional Neural Networks (CNN), provide relatively accurate estimates of the PMI. Recent studies focusing on the thanatomicrobiome are emerging as a promising tool in the forensic field, as this microbiome is unique and individualizing. These characteristics render it useful in the various stages of human identification and geolocation in criminal investigations. However, the review underscores the need for studies with larger sample sizes and for exploring the role of microorganisms beyond bacteria, in order to broaden and enhance the research landscape in this emerging field.