Trends in forensic microbiology: From classical methods to deep learning.

Area of Science:

• Forensic Science
• Microbiology
• Bioinformatics
• Artificial Intelligence

Background:

• Forensic microbiology is crucial for death investigation, individual identification, crime scene detection, and postmortem interval estimation.
• Traditional microbial culture methods are inefficient, costly, and lack quantitative analysis capabilities.
• Advancements in high-throughput sequencing, bioinformatics, and artificial intelligence are transforming the field.

Purpose of the Study:

• To review the applications and development of forensic microbiology.
• To summarize research progress in machine learning (ML) and deep learning (DL) for microbial genome sequencing and image analysis in forensics.
• To provide a future outlook on forensic microbiology.

Main Methods:

• Review of machine learning models (e.g., RF, SVM, ANN, DNN, regression, PLS, ANOSIM, ANOVA) applied to microbiome and metagenomic studies.
• Analysis of deep learning models, including convolutional neural networks (CNNs), for microorganism image analysis.
• Integration of microbial genome sequencing data with ML/DL approaches.

Main Results:

• ML and DL models show significant advancements in forensic applications compared to traditional methods.
• Deep learning models, particularly CNNs, enhance forensic prognosis accuracy through object detection in microorganism images.
• These advanced techniques offer improved efficiency and quantitative analysis in forensic microbiology.

Conclusions:

• Machine learning and deep learning are revolutionizing forensic microbiology by providing powerful analytical tools.
• The integration of genomic and imaging data with AI offers unprecedented potential for forensic investigations.
• Future research should focus on further developing and validating these advanced computational methods in forensic contexts.