Forensic science has undergone remarkable advancements over the past few decades, largely driven by the integration of molecular and biochemical techniques into criminal investigations.¹ Traditional forensic methods such as fingerprint analysis, blood typing, and DNA profiling have significantly improved the identification of individuals involved in criminal activities.² However, one of the persistent challenges in forensic investigations is the accurate determination of the time at which biological evidence was deposited at a crime scene. While DNA analysis can establish the source of biological material with high precision, it provides little information regarding the age of the biological trace, which is often crucial for reconstructing criminal events and establishing timelines.

Biochemistry offers promising solutions to this challenge through the analysis of dynamic molecular changes that occur in biological samples after deposition.³ Human sweat is a readily available biological fluid composed of water, electrolytes, amino acids, organic acids, proteins, and metabolic by-products.⁴ Following deposition on a surface, these biochemical constituents undergo progressive degradation due to environmental exposure, oxidation, microbial activity, and physicochemical reactions. Such changes occur in a relatively predictable manner and may serve as biochemical markers for estimating the age of forensic evidence.

Recent advances in metabolomics and analytical biochemistry have enabled the sensitive detection and quantification of small-molecule metabolites present in biological fluids.⁵ Techniques such as liquid chromatography–mass spectrometry (LC-MS) have facilitated comprehensive metabolic profiling, allowing investigators to monitor temporal changes in metabolite concentrations^.6 These developments have created new opportunities for forensic scientists to move beyond simple identification and toward the determination of evidence chronology. The concept of using metabolite degradation patterns as a molecular clock represents an innovative approach with significant potential for forensic applications.

Among the metabolites present in sweat, compounds such as lactate, pyruvate, creatinine, and amino acids exhibit measurable changes over time following environmental exposure^.7 Monitoring these changes may provide valuable information regarding the duration for which a biological trace has been present at a crime scene.^8,9 Furthermore, combining multiple metabolites into a composite index may improve predictive accuracy and reduce variability associated with individual biomarkers.^10,11

The present study introduces MetaboTrace, a novel forensic-biochemistry approach designed to estimate the deposition time of sweat traces through quantitative analysis of metabolite degradation patterns. By developing a Forensic Metabolic Decay Index (FMDI) and evaluating its relationship with exposure duration, this study aims to establish a reliable biochemical framework for temporal analysis of forensic evidence. The findings may contribute to the development of next-generation forensic tools capable of enhancing crime scene reconstruction and improving the interpretation of biological evidence in criminal investigations.

This prospective experimental forensic-biochemistry study was conducted on 82 healthy volunteers aged between 20 and 45 years to investigate the potential of sweat metabolite degradation patterns in estimating the age of biological traces. Individuals with chronic metabolic disorders, active dermatological conditions, recent infections, or medications known to affect sweat composition were excluded from participation. After obtaining informed consent, sweat samples were collected under standardized laboratory conditions. Participants were instructed to avoid strenuous exercise, alcohol, and caffeine consumption for 24 hours prior to sample collection. Sweat was obtained from the forearm using sterile absorbent pads following skin cleansing and transferred into sterile collection tubes for analysis. To simulate forensic biological traces, aliquots of 100 μL of sweat from each participant were deposited onto sterile glass slides and maintained in a controlled environmental chamber at a temperature of 25 ± 2°C and relative humidity of 55 ± 5%. Samples were analyzed at predetermined time intervals of 0, 6, 12, 24, 48, and 72 hours after deposition to assess temporal biochemical changes. Selected metabolites including lactate, pyruvate, urea, creatinine, glycine, alanine, and serine were quantified using liquid chromatography–mass spectrometry (LC-MS). Prior to analysis, sweat samples underwent methanol-based extraction, centrifugation, and filtration to ensure accurate metabolite detection and quantification. A novel Forensic Metabolic Decay Index (FMDI) was developed by integrating normalized concentrations of the measured metabolites into a composite score representing the degree of biochemical degradation over time. The index was designed to decrease progressively as exposure duration increased. To evaluate the predictive capability of metabolite degradation patterns, machine-learning algorithms, including Random Forest, Support Vector Machine, and Gradient Boosting models, were trained using the metabolomic dataset to estimate the deposition time of sweat traces. Statistical analysis was performed using SPSS version 27.0, with results expressed as mean ± standard deviation. Differences among time intervals were analyzed using one-way ANOVA followed by Tukey’s post hoc test, while Pearson correlation analysis was used to assess associations between metabolite concentrations, FMDI values, and exposure duration. A p-value less than 0.05 was considered statistically significant throughout the study.

A total of 82 sweat samples were analyzed over a 72-hour exposure period. Significant time-dependent degradation was observed in all measured metabolites. Lactate, pyruvate, urea, and creatinine concentrations progressively declined with increasing exposure duration, indicating continuous biochemical decomposition following deposition. The novel Forensic Metabolic Decay Index (FMDI) demonstrated a marked reduction over time and showed a strong negative correlation with exposure duration (r = –0.91, p < 0.001). Machine-learning-based prediction models successfully estimated sweat trace age with high accuracy, particularly during the early post-deposition period.

Table 1. Changes in Sweat Metabolite Concentrations at Different Exposure Times (n = 82)

Table 2. Forensic Metabolic Decay Index (FMDI) According to Exposure Duration

Table 3. Performance of the MetaboTrace Machine-Learning Model in Estimating Sweat Trace Age

The results demonstrate a statistically significant decline in sweat metabolite concentrations and FMDI scores with increasing exposure time. The strong inverse relationship between FMDI and trace age supports the utility of metabolomic degradation profiling as a biochemical marker of evidence age. Furthermore, the MetaboTrace predictive model achieved an overall accuracy of 90.6%, highlighting its potential application in forensic investigations for estimating the deposition time of biological traces.

The present study introduces MetaboTrace, a novel forensic-biochemistry approach designed to estimate the age of biological traces through the analysis of sweat metabolite degradation patterns. The findings demonstrate that metabolite concentrations in sweat decrease in a predictable and time-dependent manner following deposition, supporting the hypothesis that biochemical changes can function as a molecular clock for forensic investigations. Unlike conventional DNA-based forensic methods, which primarily provide information regarding the source of biological evidence, the proposed metabolomic approach offers valuable temporal information that may assist investigators in reconstructing crime scene events and establishing timelines. A significant decline was observed in the concentrations of lactate, pyruvate, urea, and creatinine over the 72-hour exposure period. Lactate exhibited the most pronounced reduction, decreasing from 18.4 ± 1.5 mmol/L at baseline to 3.3 ± 0.7 mmol/L after 72 hours. This progressive decline may be attributed to environmental oxidation, microbial metabolism, enzymatic degradation, and physicochemical instability following exposure to ambient conditions. Similar degradation trends were observed for pyruvate and other metabolites, suggesting that sweat contains multiple biochemical markers capable of reflecting trace age. The use of several metabolites rather than a single biomarker enhances reliability by reducing variability associated with individual metabolic pathways. One of the major innovations of this study was the development of the Forensic Metabolic Decay Index (FMDI). The index integrated multiple metabolite measurements into a single quantitative parameter, allowing a more comprehensive assessment of biological trace aging. The FMDI demonstrated a strong negative correlation with exposure duration (r = −0.91, p < 0.001), indicating that the index accurately reflected the progression of biochemical degradation. This finding suggests that composite metabolomic indicators may provide greater predictive power than individual metabolite measurements alone and could serve as a practical tool for forensic laboratories. The machine-learning component of the MetaboTrace system further strengthened the predictive capability of the approach. The overall model accuracy of 90.6% indicates that computational analysis of metabolomic data can effectively estimate deposition time. Notably, the highest prediction accuracy (95.2%) was observed during the first 12 hours after deposition, a period that is often critical in criminal investigations. The slight decline in accuracy at later time intervals may be explained by increasing environmental variability and the convergence of metabolite concentrations as degradation progresses. Nevertheless, the model maintained strong predictive performance throughout the study period, highlighting the value of integrating artificial intelligence with forensic biochemistry. The practical implications of these findings are substantial. Accurate estimation of biological trace age could help distinguish whether evidence was deposited during the commission of a crime or was present beforehand. Such information may strengthen evidentiary interpretation, support witness statements, and improve crime scene reconstruction. Furthermore, because sweat traces are commonly encountered on surfaces, clothing, and personal belongings, the MetaboTrace approach may have broad applicability across various forensic contexts. Despite its promising results, several limitations should be considered. The study was conducted under controlled environmental conditions, whereas actual crime scenes are subject to fluctuations in temperature, humidity, ultraviolet radiation, and microbial contamination that may influence metabolite degradation rates. In addition, only healthy adults were included, and individual variations related to age, diet, disease status, and medication use were not examined. Future studies should validate the proposed model under diverse environmental conditions, investigate additional metabolomic biomarkers, and evaluate its applicability to other biological fluids such as saliva, blood, and urine. Overall, the findings suggest that sweat metabolomics represents a promising frontier in forensic science. By combining biochemical analysis, metabolomic profiling, and machine-learning algorithms, the MetaboTrace system provides a scientifically robust framework for estimating biological trace exposure time. This approach has the potential to complement traditional forensic methods and contribute to the development of next-generation tools for crime scene investigation and forensic evidence interpretation.

In conclusion, the present study demonstrates that sweat metabolite degradation patterns can serve as reliable biochemical indicators for estimating the age of forensic biological traces. The novel MetaboTrace approach, incorporating the Forensic Metabolic Decay Index (FMDI) and machine-learning analysis, showed a strong association between metabolite changes and exposure duration, achieving high predictive accuracy. These findings suggest that metabolomic profiling of sweat may provide valuable temporal information that complements conventional DNA-based forensic investigations, thereby enhancing crime scene reconstruction and evidentiary interpretation. Further validation under real-world forensic conditions is warranted to establish its practical applicability.

1. Procopio N, Bonicelli A. From flesh to bones: Multi‐omics approaches in forensic science. Proteomics. 2024 Jun;24(12-13):2200335.
2. Alketbi SK. The role of DNA in forensic science: A comprehensive review. International Journal of Science and Research Archive. 2023 Aug 31;9(02):814-29.
3. Broza YY, Zhou X, Yuan M, Qu D, Zheng Y, Vishinkin R, Khatib M, Wu W, Haick H. Disease detection with molecular biomarkers: from chemistry of body fluids to nature-inspired chemical sensors. Chemical reviews. 2019 Nov 15;119(22):11761-817.
4. Liu Y, Luo T, Ding C, Xuan L, Li J, Yang R. Progress in the application of flexible and wearable electrochemical sensors in monitoring biomarkers of athletes. Advanced Materials Technologies. 2024 Dec;9(23):2400619.
5. Leong SX, Leong YX, Koh CS, Tan EX, Nguyen LB, Chen JR, Chong C, Pang DW, Sim HY, Liang X, Tan NS. Emerging nanosensor platforms and machine learning strategies toward rapid, point-of-need small-molecule metabolite detection and monitoring. Chemical Science. 2022;13(37):11009-29.
6. Ren JL, Zhang AH, Kong L, Wang XJ. Advances in mass spectrometry-based metabolomics for investigation of metabolites. RSC advances. 2018;8(40):22335-50.
7. Gnanasigamani M, Arslan E, Soylu Y, Kilit B, Chmura P. Mapping multi-modal fatigue in elite soccer through sweat-omics perspectives: A narrative review. Biology. 2025 Aug 16;14(8):1069.
8. Hossain PS, Marasco E. Sweat Biomarkers: Emerging Technologies for Collection and Comprehensive Metabolite Profiling. Authorea Preprints. 2025 Oct 17.
9. Morgan RM, Meakin GE, French JC, Nakhaeizadeh S. Crime reconstruction and the role of trace materials from crime scene to court. Wiley Interdisciplinary Reviews: Forensic Science. 2020 Jan;2(1):e1364.
10. Lewis JE, Kemp ML. Integration of machine learning and genome-scale metabolic modeling identifies multi-omics biomarkers for radiation resistance. Nature communications. 2021 May 11;12(1):2700.
11. Srivastava A, Creek DJ. Discovery and validation of clinical biomarkers of cancer: a review combining metabolomics and proteomics. Proteomics. 2019 May;19(10):1700448.