Estimating bloodstain age using headspace gas chromatography-ion mobility spectrometry (HS-GC-IMS) and machine learning.

Bloodstain age, or the time since deposition (TSD), is crucial for reconstructing event timelines in forensic investigations. While traditional bloodstain analysis has limitations in reliability and methodology, Headspace-Gas Chromatography-Ion Mobility Spectrometry (HS-GC-IMS) and machine learning (ML) provide a rapid, reliable, objective, and non-destructive solution. This study combined HS-GC-IMS with ML to systematically investigate the dynamic changes of volatile organic compounds (VOCs) during the early, middle, and late stages of bloodstain aging. Blood samples from 10 healthy volunteers were analyzed at 8 specific time points spanning 0 h to 70 days. HS-GC-IMS was performed to profile VOC, while multiple ML algorithms were applied to identify key VOC markers. Among 52 VOCs detected, we found specific changes in 22 key biomarkers—including ketones, aldehydes, and terpenes—that correlated with the early (decrease in ketone), middle (emergence of terpenes and aromatic hydrocarbons), and late (dominance of aldehydes and ketones) stages of bloodstain aging. Further OPLS-DA and Pearson correlation analysis identified five independent metabolites for estimating bloodstain age: Propanal, 2-Methoxy-3-methylpyrazine, Benzaldehyde, Ethyl Acetate, and 2,3-Pentanedione. Subsequent ML models (Random Forest, Naive Bayes, and Neural Network) exhibited strong classification performance, enabling effective discrimination of bloodstain age. The integrated HS-GC-IMS and ML approach provides a highly sensitive, rapid, and reliable way to screen for and estimate bloodstain age at crime scenes.