Key points are not available for this paper at this time.
Body mass index (BMI) is a crucial phenotypic feature with significant application value in forensic investigations; however, inferring BMI from forensic-related biological samples such as saliva remains challenging. In this study, saliva samples were collected and subjected to 16S rRNA sequencing to characterize microbial community composition, and BMI-associated microbial markers were screened using linear discriminant analysis effect size and the Kruskal-Wallis rank-sum test. A random forest model was subsequently constructed to infer BMI categories based on the selected microbial markers. The results showed that twenty-two microbial taxa, including the genera Neisseria, Veillonella, Prevotella, Streptococcus, and Achromobacter, exhibited significantly different abundance distributions among BMI groups and could serve as BMI-associated microbial indicators. Principal coordinates analysis demonstrated a clear separation between normal-weight and overweight groups, and the random forest model accurately inferred BMI categories for most samples. These findings indicate that saliva-associated microbial markers have potential as valuable indicators for BMI inference, providing a promising tool for forensic research.
Ji et al. (Tue,) studied this question.