Bloodstain leaving time detection method

Abstract

The invention provides a bloodstain leaving time detection method, and belongs to the technical field of forensic medicine detection. The method comprises the following steps: sampling a carrier attached with bloodstains to obtain a first sample, and sampling a carrier not attached with bloodstains to obtain a second sample; preparing an internal standard solution with a set concentration, and respectively filling the internal standard solutions with the same volume into different headspace bottles; respectively combining the first sample and the second sample with a headspace bottle, and respectively putting into a headspace solid-phase extractor for extraction treatment; analyzing by using a gas chromatograph-mass spectrometer to obtain a volatile organic compound (VOC) map, and determining actually measured VOC components according to the VOC map corresponding to the first sample and the VOC map corresponding to the second sample; and determining the bloodstain leaving time according to the actually measured VOC components and the bloodstain leaving time analysis model. According to the method, the influence of object factors on the bloodstain leaving time inference result can be reduced, and the accuracy and efficiency of the bloodstain leaving time inference result are improved.

Description

Technical Field

[0001] The present application relates to the field of forensic detection technology, and in particular to a method for detecting the time a blood stain remains. Background Art

[0002] Bloodstains are one of the most common biological evidences at all types of crime scenes. As an objective existence at the crime scene, they contain extremely rich criminal information and are important evidence for judging the nature of the case, analyzing the crime process, determining the direction of investigation, and proving the crime. At present, the research on bloodstains mainly focuses on the following two aspects: one is to reconstruct the case by analyzing the morphology of bloodstains, and the other is to extract DNA from bloodstains for individual identification. However, it is still difficult to infer the time of the crime through bloodstains.

[0003] At present, the inference of bloodstains’ time mainly relies on observing the shape and color of bloodstains and making rough empirical inferences at the scene, but there is a lack of sufficient quantitative basis, which can easily lead to large errors. Laboratory testing attempts to make inferences from the perspective of chemical composition or biology, and there are research reports on the use of RNA analysis, electron paramagnetic resonance, single cell gel electrophoresis and other methods to infer the time of bloodstains. However, these methods have strict testing conditions, high professional requirements for the examiner, and often damage the shape of bloodstains, which are not suitable for rapid time inference of trace bloodstains at the scene. With the development of instrumental analysis methods such as spectroscopy technology, some researchers have used liquid chromatography-mass spectrometry, near-infrared spectroscopy, ultraviolet-visible spectroscopy, reflectance spectroscopy and other technologies to infer the time of bloodstains.

[0004] In the process of implementing the embodiments of the present application, it is found that there are at least the following technical problems in the related technology:

[0005] The existing schemes for inferring the time when bloodstains were left are greatly affected by objective factors, and the accuracy of the results of inferring the time when bloodstains were left needs to be further improved. Summary of the invention

[0006] The embodiment of the present application provides a bloodstain remaining time detection method to solve the problem that objective factors affect the accuracy of bloodstain remaining time inference results.

[0007] The present application embodiment provides a bloodstain remaining time detection method, comprising:

[0008] Sampling the carrier with blood stains to obtain a first sample, and sampling the carrier without blood stains to obtain a second sample;

[0009] Prepare an internal standard solution of set concentration, and take the same volume of the internal standard solution and put it into different headspace bottles respectively;

[0010] The first sample and the second sample are respectively combined with headspace bottles, and are respectively placed in headspace solid phase extractors for extraction treatment;

[0011] Using a gas chromatography-mass spectrometer to analyze, obtain a volatile organic compound (VOC) spectrum, and determine the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample;

[0012] The bloodstain retention time is determined based on the measured VOC components and the bloodstain retention time analysis model.

[0013] In a possible implementation, determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample includes:

[0014] performing differential processing on the VOC spectrum corresponding to the first sample based on the VOC spectrum corresponding to the second sample to obtain a differentially processed spectrum;

[0015] The measured VOC composition is determined based on the differentially processed map.

[0016] In a possible implementation, performing differential processing on the VOC spectrum corresponding to the first sample based on the VOC spectrum corresponding to the second sample to obtain a differentially processed spectrum includes:

[0017] Analyzing the VOC spectrum corresponding to the second sample to determine a first characteristic peak set, and analyzing the VOC spectrum corresponding to the first sample to determine a second characteristic peak set;

[0018] Based on the second characteristic peak set, the characteristic data corresponding to the first characteristic peak set is deducted to obtain a differential processing spectrum.

[0019] In a possible implementation, the measured VOC composition includes a VOC composition of a set difference substance;

[0020] Before determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample, the method further includes:

[0021] Acquire a historical blood sample of the blood type corresponding to the first test sample;

[0022] The VOC content of the historical blood sample is sampled at set time intervals, and an orthogonal partial least squares discriminant analysis (OPLS-DA) is performed to obtain the variable importance in projection (VIP) value and the correlation coefficient p value between the predicted sample score and the independent variable;

[0023] According to the VIP value and the p value of the correlation coefficient between the predicted sample score and the independent variable, the difference substance type characterizing the time difference is screened out as the difference substance.

[0024] In a possible implementation, the storage temperature of the historical blood sample is 25°C.

[0025] In a possible implementation, determining the bloodstain duration according to the measured VOC components and the bloodstain duration analysis model includes:

[0026] Input the measured VOC components into multiple bloodstain residual time analysis models respectively to obtain multiple bloodstain residual times; wherein each bloodstain residual time analysis model is constructed based on one or more of DT, SVM, Lasso, KNN, XGBoost, and RF regression models;

[0027] The final bloodstain remaining time is determined based on the respective bloodstain remaining times.

[0028] In a possible implementation, determining the final bloodstain remaining time based on each bloodstain remaining time includes:

[0029] Calculate the mean of the bloodstain remaining time, and use the mean calculation result as the final bloodstain remaining time; or,

[0030] According to the distribution of the time when each bloodstain was left, one is determined as the final time when the bloodstain was left.

[0031] In a possible implementation, the internal standard solution is a bromobenzene solution; and the set concentration is 20 ppb.

[0032] In a possible implementation, the internal standard solution further includes chromatographically pure methanol solvent; and the volume of the internal standard solution is 10 μl.

[0033] In a possible implementation, the extraction temperature corresponding to the extraction process is 50° C., and the extraction time is 45 minutes.

[0034] In a possible implementation, the gas chromatography-mass spectrometry analysis corresponding conditions include:

[0035] Chromatographic conditions: HP-5MS capillary column (60m×0.25mm×1.0μm); programmed temperature conditions: initial column temperature 40°C, hold for 1 min, increase to 80°C at 5°C / min, hold for 1 min, then increase to 260°C at 20°C / min, hold for 5 min; carrier gas is high-purity helium (99.999%), constant flow mode, flow rate is 1.0mL / min; splitless injection, solvent delay 3min; injection port temperature is 280°C; vacuum compensation and carrier gas saving are turned on, flow rate is 20.0mL / min, hold for 2min; mass spectrometry conditions: electron bombardment ion source (EI), electron energy is 70eV, transmission line temperature is 280°C, ion source temperature is 280°C; scanning mass range: 40-450amu, scanning time: 0.2s; filament current: 20μA.

[0036] The embodiment of the present application provides a method for detecting the residual time of bloodstains, by sampling carriers with and without bloodstains, respectively, and combining the collected samples with internal standard solutions, and directly placing them in a headspace solid phase extractor for extraction treatment. No anticoagulant is used in the process, thereby simplifying the preliminary processing steps of bloodstain analysis. A gas chromatography-mass spectrometer is used for analysis to obtain a VOC spectrum, and then the measured VOC components are determined according to the VOC spectrum corresponding to the first sample with attached bloodstains and the VOC spectrum corresponding to the second sample without attached bloodstains, and the bloodstain residual time is determined according to the measured VOC components and the bloodstain residual time analysis model. Among them, based on the non-targeted volatile omics method, headspace solid phase microextraction-gas chromatography-mass spectrometry technology is used to quickly obtain blood VOC information, and the VOC spectrum corresponding to the carrier sample with attached bloodstains is corrected with the spectrum analysis results of the carrier without attached bloodstains, which can reduce the influence of objective factors on the bloodstain residual time inference results and improve the accuracy and efficiency of the bloodstain residual time inference results. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0038] Figure 1 This is a flow chart of a method for detecting the residual time of bloodstains provided in an embodiment of the present application;

[0039] Figure 2 This is a sample score graph of the OPLS-DA model of blood for 7 days provided in a specific embodiment of the present application;

[0040] Figure 3 is a schematic diagram of a permutation test result provided by a specific embodiment of the present application;

[0041] Figure 4 This is a schematic diagram of potential difference substances of blood VOC provided by a specific embodiment of the present application;

[0042] Figure 5 This is a result diagram of the difference between the predicted values ​​and the actual values ​​of various models provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0044] The terms "first", "second", etc. in the specification and claims of the embodiments of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the embodiments of the present application described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0045] Unless otherwise specified, the term "plurality" means two or more. The character " / " indicates that the objects before and after are in an "or" relationship. For example, A / B means: A or B. The term "and / or" is a description of the association relationship between objects, indicating that there can be three relationships. For example, A and / or B means: A or B, or, A and B.

[0046] The words used in this application are only used to describe the embodiments and are not used to limit the claims. As used in the description of the embodiments and claims, unless the context clearly indicates, the singular forms of "a", "an" and "the" are intended to include the plural forms as well. Similarly, the term "and / or" as used in this application refers to any and all possible combinations of one or more associated listings. In addition, when used in this application, the term "comprise" and its variants "comprises" and / or including (comprising) refer to the existence of stated features, wholes, steps, operations, elements, and / or components, but do not exclude the existence or addition of one or more other features, wholes, steps, operations, elements, components and / or these groups. In the absence of more restrictions, the elements defined by the statement "including one..." do not exclude the existence of other identical elements in the process, method or device including the elements.

[0047] In this application, each embodiment may focus on the differences from other embodiments, and the same or similar parts between the embodiments may refer to each other. For the methods, products, etc. disclosed in the embodiments, if they correspond to the method part disclosed in the embodiments, then the relevant parts may refer to the description of the method part.

[0048] The current scheme for inferring the time of bloodstains has disadvantages such as subjectivity, complexity, and destructiveness to the samples. In order to facilitate the actual work of the public security, this application has designed a new method based on accuracy. This method uses headspace solid phase microextraction technology for pre-treatment, and based on the principle of volatile omics, uses gas chromatography-mass spectrometry to analyze the content of volatile organic compounds in bloodstains. Through machine learning, the time of bloodstains is inferred based on the changes in the content of time markers in volatile organic compounds in the blood, and the time of bloodstains leaving the body is inferred more accurately. It is suitable for the inference of the time of dried blood, has a wide range of applications, strong anti-interference ability, and strong result interpretation ability, and provides more powerful evidence for judging the nature of the case, analyzing the process of committing the crime, determining the direction of investigation, and proving the crime.

[0049] In order to make the purpose, technical solutions and advantages of the present application clearer, specific embodiments will be described below in conjunction with the accompanying drawings.

[0050] Figure 1 FIG. 1 is a flow chart of a method for detecting the residual time of bloodstains provided in an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0051] S101, sampling the carrier with blood stains to obtain a first sample, and sampling the carrier without blood stains to obtain a second sample.

[0052] At different crime scenes, bloodstains are attached to different carriers. After a series of sampling, extraction, and VOC component analysis, the carrier components will affect the VOC component analysis results of bloodstains. To avoid this problem, when sampling at the scene, the carrier without bloodstains is sampled at the same time. Extraction and VOC component analysis are performed based on the carrier sampling results to improve the accuracy of bloodstain retention time estimation.

[0053] In the actual implementation process, in order to further improve the accuracy of bloodstain retention time estimation, the volumes corresponding to the first sample and the second sample are the same.

[0054] S102, preparing an internal standard solution of a set concentration, and taking the same volume of the internal standard solution and respectively filling it into different headspace bottles.

[0055] S103, respectively combining the first sample and the second sample with the head space bottle, and respectively placing them in a head space solid phase extractor for extraction treatment.

[0056] Among them, the internal standard solution of set concentration is prepared, and the same volume of internal standard solution is taken and loaded into different headspace bottles to meet the quantitative analysis requirements. The same volume ensures the accuracy and reliability of the analysis results.

[0057] The first sample and the second sample are combined with the headspace bottle, and the first sample and the headspace bottle combination are placed in a headspace solid phase extractor for extraction. After that, the first sample and the headspace bottle combination are taken out, and the second sample and the headspace bottle combination are placed in a headspace solid phase extractor for extraction. The experimental results of the second sample are used as reference data or corrected data for the first sample. Based on the experimental results of the second sample, the VOC components of the carrier attached with blood stains are prevented from affecting the detection results of the VOC components in the blood.

[0058] S104, using a gas chromatography-mass spectrometer to analyze and obtain a VOC spectrum, and determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample.

[0059] Among them, a gas chromatography-mass spectrometer was used to analyze the samples, and the first sample and the second sample were analyzed respectively to obtain the corresponding VOC spectra. The VOC spectrum corresponding to the carrier sample with blood stains was corrected with the spectrum analysis results of the carrier without blood stains, which can reduce the influence of the objective factors on the inference results of the blood stain retention time.

[0060] The embodiment of the present application is based on volatile omics, and the required pre-treatment is simple, which improves the shortcomings of the existing bloodstain residual time estimation that is easily affected by object interference and has poor stability.

[0061] S105, determining the bloodstain retention time according to the measured VOC components and the bloodstain retention time analysis model.

[0062] Among them, the bloodstain time analysis model is built based on a machine learning algorithm, which can improve the efficiency and accuracy of bloodstain time inference.

[0063] In this embodiment, the carriers with and without blood stains are sampled respectively, and the collected samples are combined with internal standard liquids respectively, and directly placed in a headspace solid phase extractor for extraction treatment. No anticoagulant is used in the process, thereby simplifying the preliminary processing steps of blood stain analysis and being closer to the blood sample situation in actual cases. Gas chromatography-mass spectrometry is used for analysis to obtain a VOC spectrum, and then the measured VOC components are determined according to the VOC spectrum corresponding to the first sample with blood stains and the VOC spectrum corresponding to the second sample without blood stains, and the blood stain retention time is determined according to the measured VOC components and the blood stain retention time analysis model. Among them, based on the non-targeted volatile omics method, headspace solid phase microextraction-gas chromatography-mass spectrometry technology can quickly obtain blood VOC information, and the VOC spectrum corresponding to the carrier sample with blood stains attached is corrected by the spectrum analysis results of the carrier without blood stains attached, which can reduce the influence of objective factors on the blood stain retention time inference results and improve the accuracy and efficiency of the blood stain retention time inference results.

[0064] In different embodiments, the manner of determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample is different.

[0065] In a possible implementation, in step S104, determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample includes:

[0066] Based on the VOC spectrum corresponding to the second sample as control data,

[0067] When a characteristic peak different from the control data appears in the VOC spectrum corresponding to the first sample or the peak intensity change is greater than the set value, the compound corresponding to the characteristic peak or the peak intensity change greater than the set value is determined to be the measured VOC component.

[0068] In this embodiment, only the VOC spectrum corresponding to the second sample is used as the control data, and the VOC spectrum corresponding to the first sample is not differentiated, which simplifies the data processing process, reduces the data processing volume, and is suitable for rapid screening or preliminary analysis of the measured VOC composition.

[0069] In another possible implementation, in step S104, determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample includes:

[0070] performing differential processing on the VOC spectrum corresponding to the first sample based on the VOC spectrum corresponding to the second sample to obtain a differentially processed spectrum;

[0071] The measured VOC composition is determined based on the differentially processed spectrum.

[0072] In one possible implementation, performing differential processing on the VOC spectrum corresponding to the first sample based on the VOC spectrum corresponding to the second sample to obtain the differentially processed spectrum includes:

[0073] Analyzing the VOC spectrum corresponding to the second sample to determine the first characteristic peak set, and analyzing the VOC spectrum corresponding to the first sample to determine the second characteristic peak set;

[0074] Based on the second characteristic peak set, the characteristic data corresponding to the first characteristic peak set is deducted to obtain a differential processing spectrum.

[0075] Among them, based on the second characteristic peak set, the characteristic data corresponding to the first characteristic peak set is deducted,

[0076] This can be done automatically by software, by subtracting the overlapping parts of the two sets of data, retaining the characteristic peaks that only appear in the first sample, and obtaining a differentially processed spectrum.

[0077] In this embodiment, the VOC spectrum corresponding to the first sample is differentiated based on the VOC spectrum corresponding to the second sample, which improves the accuracy of data processing and can more accurately identify the unique VOC components in the first sample, thereby providing more authoritative evidence for criminal on-site investigation and avoiding misjudgment.

[0078] In order to further improve the efficiency of bloodstain retention time detection, the types of VOC components analyzed by gas chromatography-mass spectrometry analysis can be limited to improve the data targeting during the analysis and processing process, thereby reducing the amount of data processing and improving the analysis and processing efficiency.

[0079] In one possible implementation, the measured VOC composition includes setting the VOC composition of the difference substance;

[0080] Before determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample, the method further includes:

[0081] Obtain a historical blood sample of the blood type corresponding to the first test sample;

[0082] The VOC content of historical blood samples was sampled at set time intervals and analyzed by OPLS-DA to obtain the VIP value and the correlation coefficient p value between the predicted sample score and the independent variable;

[0083] According to the VIP value and the p-value of the correlation coefficient between the predicted sample score and the independent variable, the differential substance types characterizing the time difference are screened out as differential substances.

[0084] In actual implementation, there are many types of blood corresponding to various crime scenes. Specifically, blood is divided into human blood and non-human blood.

[0085] When the blood is human blood, the age, disease, gender, etc. are different, and the test results corresponding to the same blood stain retention time may be different. For example: as people age, the coagulation function of the human body will gradually decline, so the test results corresponding to the same blood stain retention time of blood from people of different age groups may be different; hemophilia is a hereditary coagulation disorder. If the blood left at the crime scene is the blood of a hemophiliac, the test results corresponding to the same blood stain retention time may be different from those of ordinary people.

[0086] When the blood is not human blood, the species, life span, etc. are different, and the test results corresponding to the same blood stain retention time may be different. For example: in daily life, the test results corresponding to the same blood stain retention time of poultry (chickens, ducks, geese, etc.) and livestock (pigs, cattle, sheep, etc.) may be different, and the test results corresponding to the same blood stain retention time of animals living in the wild may also be different.

[0087] Optionally, the blood type corresponding to the first sample includes human blood and non-human blood. Further, the blood type classification corresponding to human blood includes ordinary human blood and human blood with coagulation dysfunction. The blood type classification corresponding to non-human blood includes poultry blood, livestock blood and wild creature blood.

[0088] The test results of bloodstain retention time are crucial for judging the nature of the case, analyzing the crime process, determining the direction of investigation, proving the crime and the subsequent sentencing process. Based on the above introduction, it can be seen that there are many factors that affect the test results. Therefore, after determining the measured VOC components based on the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample, determining the corresponding potential difference substances based on the blood type at the crime scene, and using them as blood time markers, the accuracy of the bloodstain retention time inference results can be further improved.

[0089] Among them, OPLS-DA is a method based on orthogonal partial least squares regression for discrimination. It makes predictions and classifications by finding linear combinations to maximize the correlation between variables and categories. After processing the data with OPLS-DA, the VIP value can be obtained by calculating the weight of each variable on each component and the weight of the component's contribution to the overall model. The higher the VIP value, the more important the component is.

[0090] According to the VIP value and the p-value of the correlation coefficient between the predicted sample score and the independent variable, the differential substance types that characterize the time difference are screened out as differential substances. When determining the bloodstain time based on the measured VOC composition and the bloodstain time analysis model, the attention paid to the differential substances is increased, thereby improving the accuracy of the bloodstain time analysis results.

[0091] The embodiments of the present application focus on human blood samples as an example for illustration. Among them, the blood samples are all collected from veins. In order to simulate the crime scene, no anticoagulant is used for anticoagulation after blood collection.

[0092] In a specific implementation, the process of determining the difference substance type characterizing the time difference based on the historical blood samples is as follows:

[0093] 1. Sample collection

[0094] A total of 56 blood samples were collected from the arms of 8 unrelated individuals using 10ml vacuum blood collection tubes without additives. 300μl of each blood sample was placed on a sterile blood collection card and dried in a dark and ventilated place. The dried blood card was stored at 25°C. Among them, the 8 unrelated individuals were aged between 23 and 28 years old.

[0095] 2. Blood VOC data acquisition

[0096] Headspace solid phase microextraction was used as a pretreatment method every 24 hours, and GC-MS analysis was performed with bromobenzene as the internal standard. The NIST 17 spectral library was used for qualitative retrieval and the internal standard method for quantitative analysis. The metabolic public database HMDB was used for comparison and matching to exclude non-endogenous substances, and qualitative and quantitative data of various VOCs at different times within 7 days were obtained.

[0097] 3. Screening of time markers

[0098] The acquired VOC data were analyzed by OPLS-DA, a model was established, and the model was subjected to 200 permutation tests. The p-value of the correlation coefficient between the predicted sample score and the independent variable was calculated, and the VIP value was calculated. Potential differential substances were screened when the VIP was greater than 0.8 and the p-value of the correlation coefficient between the predicted sample score and the independent variable was less than 0.05.

[0099] Figure 2 The OPLS-DA model sample score graph of blood samples taken over 7 days is shown. In the two-dimensional score graph, the different color clustering results show that the blood samples taken over 7 days are clearly divided into 7 groups, indicating that the blood VOCs show differences over the 7 days. Figure 3 This shows that the R 2 and Q 2 The values ​​are lower than the original values, indicating that the model is not overfitting. Figure 4The results showed that the potential differential substances of the model were diethylene glycol monobutyl ether, dibutyl phthalate, hexanal, n-undecane, diisobutyl phthalate, 3,3-dimethylhexane, 2,2,4-trimethylpentanediol diisobutyrate, azulene, ethylbenzene, 2,4-dimethyl-3-hexanone, N,N-dimethylbenzylamine, o-xylene, methyl heptenone, and decanal.

[0100] based on Figure 4 The analysis results limit the time markers of ordinary human blood (i.e., the set difference substances included in the measured VOC components) to the above 14 items. Other substances whose VOC components remain basically unchanged over time are no longer analyzed in order to improve the efficiency of analysis and processing.

[0101] In other embodiments, different substances are specifically determined and set as time markers for the blood of people of different age groups, the blood of people with hereditary coagulation disorders, non-human blood, etc., to improve the accuracy of bloodstain retention time detection to meet the needs of determining criminal facts such as illegal hunting, slaughtering, and animal trading.

[0102] Based on the foregoing embodiment, in a possible implementation, the storage temperature of the historical blood samples is 25°C.

[0103] In actual implementation, the storage temperature of the blood sample after leaving the body will affect the VOC spectrum of the blood sample. In this embodiment, the storage temperature of the historical blood sample is 25°C, which can simulate the environment of most crime scenes, thereby improving the authenticity of the VOC spectrum of the historical blood sample in describing the VOC information of bloodstains at different retention times.

[0104] In different embodiments, in step S105, when the bloodstain remaining time is determined according to the measured VOC components and the bloodstain remaining time analysis model, the type of the bloodstain remaining time analysis model is different.

[0105] In a possible implementation, the bloodstain retention time is determined according to the measured VOC components and the bloodstain retention time analysis model, including:

[0106] The measured VOC components are input into the bloodstain residual time analysis model constructed based on the set model to obtain the bloodstain residual time; wherein the set model is any one of DT, SVM, Lasso, KNN, XGBoost, and RF regression models.

[0107] In this embodiment, the measured VOC components are input into a bloodstain time analysis model constructed based on a set model to obtain the bloodstain time, which can improve data processing efficiency. In order to improve the accuracy of determining the bloodstain time, the set model needs to be trained and screened by combining historical blood samples to obtain multiple models.

[0108] In another possible implementation, the bloodstain retention time is determined according to the measured VOC components and the bloodstain retention time analysis model, including:

[0109] Inputting the measured VOC components into multiple bloodstain residual time analysis models respectively to obtain multiple bloodstain residual times; wherein each bloodstain residual time analysis model is constructed based on one or more of DT, SVM, Lasso, KNN, XGBoost, and RF regression models;

[0110] The final bloodstain remaining time is determined based on the respective bloodstain remaining times.

[0111] Based on this embodiment, in a possible implementation, determining the final bloodstain remaining time based on each bloodstain remaining time includes:

[0112] Calculate the mean of the bloodstain remaining time, and use the mean calculation result as the final bloodstain remaining time; or,

[0113] According to the distribution of the time when each bloodstain was left, one is determined as the final time when the bloodstain was left.

[0114] The average of each bloodstain time is calculated to improve the calculation efficiency. According to the distribution of each bloodstain time, when repeated results are selected, the bloodstain time with the largest number of repeated results is used as the final bloodstain time, or the middle value in the middle is used as the final bloodstain time.

[0115] In the embodiments of the present application, in the process of constructing the bloodstain retention time analysis model, the machine learning regression models used are: Decision Tree (DT), Support Vector Machine (Support Vector Machine), K-Nearest Neighbor (KNN), Least Absolute Shrinkage and Selection Operator (Lasso), Extreme Gradient Boosting Tree (XGBoost), and Random Forest (RF).

[0116] Among them, DT is a supervised learning algorithm that performs regression prediction by building a tree structure. It continuously divides the data set, selects the optimal partitioning feature at each node, and finally generates a decision tree to predict the output value of new data. It has the advantages of strong model structure explanation, fast construction speed, and strong robustness, but a small change in the training data will also lead to a large difference in the model results, and the risk of overfitting is high.

[0117] SVM is a regression algorithm based on the principle of structural risk minimization. It achieves regression prediction by finding a hyperplane in high-dimensional space so that the distance from all training data points to the hyperplane is minimized. It is suitable for processing high-dimensional data and small sample data, and can handle nonlinear problems. It can obtain the global optimal solution by solving convex optimization problems. However, its calculation complexity is high and the parameter selection is relatively sensitive.

[0118] KNN is an instance-based learning method that finds the k training points closest to the point to be predicted and takes the average or weighted average of the output values ​​of these k points as the prediction result. It is applicable to a wide range of data types and is insensitive to outliers. However, its calculation is more complicated and is more sensitive to sample imbalance.

[0119] Lasso is a regression analysis method widely used in statistics. Its core is to compress coefficients to achieve the purpose of variable selection and complexity adjustment, thereby improving the prediction accuracy and explanatory power of the model. It can automatically select features and effectively prevent overfitting, especially when dealing with multicollinear data or high-dimensional data. However, it is difficult to select regularization coefficients, resulting in underfitting.

[0120] XGBoost is an ensemble learning method based on the gradient boosting framework. It performs regression prediction by optimizing the combination of multiple decision trees. Since it supports custom loss functions and regularization terms, it is highly flexible and can perform parallel calculations to increase training speed. However, it requires high computing resources and is prone to overfitting.

[0121] RF is an ensemble learning method that builds multiple decision trees and takes their average to perform regression prediction. It has the advantages of strong robustness and easy parallelization. However, it has a large amount of calculation and is not suitable for small sample data.

[0122] Since the above machine learning regression models have different advantages and disadvantages, in order to avoid the influence of a single model on the bloodstain time, the embodiment of the present application combines multiple machine learning regression models to construct a bloodstain time analysis model. Specifically, the root mean square error (RMSE), the coefficient of determination (R-Square, R 2 ) and mean absolute error (MAE) were used to evaluate the model performance and select the optimal classification algorithm. After hyperparameter tuning, the optimal blood volatile organic compound-time regression model (i.e., bloodstain retention time analysis model) was obtained to achieve accurate inference of bloodstain retention time.

[0123] Before model training, blood VOC difference data over a period of time (such as blood VOC within 7 days) was obtained as sample data and divided into training set and test set in a ratio of 7:3.

[0124] Then, when setting the hyperparameters of the regression model, the DT model includes tree depth and minimum sample split number; the SVM model includes regularization parameters and kernel function type; the KNN model includes the number of neighbors and distance-weighted kernel function; the Lasso model includes the regularization strength lambda value; the XGBoost model includes the maximum tree depth, number of trees, learning rate, minimum sample split number, and minimum loss reduction required for node splitting; the RF model includes the number of trees, maximum tree depth, minimum sample split number, and number of candidate split features created by the tree. Hyperparameter tuning is performed through grid search or cross-validation.

[0125] Then, use RMSE, R 2 and MEA to measure classification model performance.

[0126] RMSE is the square root of the average of the squares of the differences between the model's predicted value and the true value. It measures the deviation between the predicted value and the actual value. The smaller the value, the closer the model's predicted result is to the true value, and the better the model performance. When measuring the performance of a classification model based on RMSE, the following formula is used:

[0127]

[0128] Among them, RMSE is the root mean square error; y i is the true value; is the predicted value; n is the number of samples.

[0129] R 2 It indicates the proportion of the variation explained by the model to the total variation and is an indicator of the goodness of fit of the model. It reflects the degree of correlation between the model prediction value and the actual value. 2 The value range is from 0 to 1. 2 The closer it is to 1, the better the model fit is and the higher the proportion of variation in the dependent variable that the independent variable can explain.

[0130] Based on R 2 When measuring the performance of the classification model, the following formula is used:

[0131]

[0132] Among them, R 2 is the coefficient of determination; is the average value of the true value; y i is the true value; is the predicted value; n is the number of samples.

[0133] MAE is the average of the absolute values ​​of the differences between the model's predicted values ​​and the true values. It directly measures the average deviation between the predicted values ​​and the actual values. The smaller the MAE value, the closer the model's predicted results are to the true values, and the better the model performance. When measuring the performance of a classification model based on MEA, the following formula is used:

[0134]

[0135] Where MAE is the mean absolute error; y i is the true value; is the predicted value; n is the number of samples.

[0136] Combined with the introduction of blood types in the above embodiments, according to historical blood samples of different blood types, the optimal model is selected from the six models to construct a bloodstain legacy time analysis model to improve the accuracy of the bloodstain legacy time inference results. In the specific implementation process, those skilled in the art can also provide other model optimization solutions to improve the model prediction accuracy without departing from the principles of the present invention. These do not make the essence of the corresponding technical solutions depart from the spirit and scope of the technical solutions of each embodiment of the present application, and should be included in the protection scope of the present application.

[0137] In a specific embodiment, the following Table 1 shows the comparison results of evaluation index values ​​of six models: DT, SVM, Lasso, KNN, XGBoost and RF. Figure 5 The following is a graph showing the difference between the predicted values ​​and the actual values ​​using six models. Figure 5 It can be seen that a variety of regression model algorithms have shown strong prediction performance, which means that the differences in blood VOC components at different time points within 7 days can be recognized and learned by the model. And according to Table 1, it can be seen that the six models have good prediction performance using RMSE, R 2 When measuring the performance of classification models with MEA, the performance results are not much different. Figure 5 The red solid line corresponds to the test set fitting line, the blue solid line corresponds to the training set fitting line, and the black solid line corresponds to the actual value. Figure 5 The following are the difference results between the predicted values ​​and the actual values ​​corresponding to the six models: RF, SVM, Lasso, DT, XGBoost and KNN.

[0138] Table 1 Evaluation index values

[0139]

[0140] Therefore, when the computing power is sufficient, when determining the bloodstain time based on the measured VOC components and the bloodstain time analysis model, you can choose to input the measured VOC components into the bloodstain time analysis model constructed based on the above six models respectively, and combine multiple bloodstain time to determine the final bloodstain time, so as to improve the calculation accuracy.

[0141] When the computing power is limited, MAE, RMSE, R 2 The bloodstain remaining time analysis model constructed by the RF model with relatively excellent comprehensive measurement results determines the final bloodstain remaining time.

[0142] The above embodiments mainly introduce the process of determining the time of blood stain. In the specific implementation process, the internal standard solution, the extraction conditions of the headspace solid phase extractor and the analysis conditions of the gas chromatography-mass spectrometry will also affect the accuracy of determining the time of blood stain.

[0143] In a possible implementation, the internal standard solution is a bromobenzene solution; the concentration is set to 20 ppb.

[0144] In a possible implementation, the internal standard solution further includes chromatographically pure methanol solvent; the volume of the internal standard solution is 10 μl.

[0145] In a possible implementation, the extraction temperature corresponding to the extraction process is 50° C., and the extraction time is 45 minutes.

[0146] In a possible implementation, the gas chromatography-mass spectrometry analyzer analyzes the corresponding conditions, including:

[0147] Chromatographic conditions: HP-5MS capillary column (60m×0.25mm×1.0μm); programmed temperature conditions: initial column temperature 40°C, hold for 1 min, increase to 80°C at 5°C / min, hold for 1 min, then increase to 260°C at 20°C / min, hold for 5 min; carrier gas is high-purity helium (99.999%), constant flow mode, flow rate is 1.0mL / min; splitless injection, solvent delay 3min; injection port temperature is 280°C; vacuum compensation and carrier gas saving are turned on, flow rate is 20.0mL / min, hold for 2min; mass spectrometry conditions: electron bombardment ion source (EI), electron energy is 70eV, transmission line temperature is 280°C, ion source temperature is 280°C; scanning mass range: 40-450amu, scanning time: 0.2s; filament current: 20μA.

[0148] The full scan mode was used to determine the retention time of each compound, the NIST 17 spectral library was used for search and comparison, and the internal standard method was used for quantification, with bromobenzene as the internal standard. Under these analytical conditions, the VOC components in the blood can be detected smoothly.

[0149] In a possible implementation, the headspace solid phase extractor is a large-volume headspace solid phase extractor, specifically, a BF-1112 large-volume headspace sampler. The solid phase microextraction needle used is "DVB-PDMS".

[0150] It should be noted that all similar replacements and modifications of process parameters are obvious to those skilled in the art, and they are deemed to be included in this application. The methods and applications of this application are described through preferred embodiments, and relevant personnel can modify, change, and combine the methods and applications of this application without departing from the content, spirit, and scope of the present invention to implement and apply the technology of the present invention.

[0151] It should be understood that the order of execution of the steps in the above embodiments does not necessarily mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0152] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A bloodstain remaining time detection method, characterized in that: include: Sampling the carrier with blood stains to obtain a first sample, and sampling the carrier without blood stains to obtain a second sample; Prepare an internal standard solution of set concentration, and take the same volume of the internal standard solution and put it into different headspace bottles respectively; The first sample and the second sample are respectively combined with headspace bottles, and are respectively placed in headspace solid phase extractors for extraction treatment; Using a gas chromatograph-mass spectrometer to analyze, obtain a volatile organic compound (VOC) spectrum, and determine the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample; The bloodstain retention time is determined based on the measured VOC components and the bloodstain retention time analysis model.

2. The method according to claim 1, characterized in that The determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample includes: performing differential processing on the VOC spectrum corresponding to the first sample based on the VOC spectrum corresponding to the second sample to obtain a differentially processed spectrum; The measured VOC composition is determined based on the differentially processed map.

3. The method according to claim 1 or 2, characterized in that: The measured VOC composition includes the VOC composition of the set difference substances; Before determining the measured VOC components according to the VOC spectrum corresponding to the first sample and the VOC spectrum corresponding to the second sample, the method further includes: Acquire a historical blood sample of the blood type corresponding to the first test sample; The VOC content of the historical blood sample is sampled at set time intervals, and an orthogonal partial least squares discriminant OPLS-DA analysis is performed to obtain a variable importance projection VIP value and a correlation coefficient p value between the predicted sample score and the independent variable; According to the VIP value and the p value of the correlation coefficient between the predicted sample score and the independent variable, the difference substance type characterizing the time difference is screened out as the difference substance.

4. The method according to claim 3, characterized in that The storage temperature of the historical blood samples was 25°C.

5. The method according to claim 1, characterized in that The determining of the bloodstain retention time according to the measured VOC components and the bloodstain retention time analysis model comprises: Input the measured VOC components into multiple bloodstain residual time analysis models respectively to obtain multiple bloodstain residual times; wherein each bloodstain residual time analysis model is constructed based on one or more of DT, SVM, Lasso, KNN, XGBoost, and RF regression models; The final bloodstain remaining time is determined based on the respective bloodstain remaining times.

6. The method according to claim 5, characterized in that The method of determining the final bloodstain remaining time based on each bloodstain remaining time includes: Calculate the mean of the bloodstain remaining time, and use the mean calculation result as the final bloodstain remaining time; or, According to the distribution of the time when each bloodstain was left, one is determined as the final time when the bloodstain was left.

7. The method according to claim 1, characterized in that The internal standard solution is bromobenzene solution; the set concentration is 20 ppb.

8. The method according to claim 7, characterized in that The internal standard solution also includes chromatographically pure methanol solvent; the volume of the internal standard solution is 10 μl.

9. The method according to claim 1, characterized in that: The extraction temperature corresponding to the extraction treatment is 50° C., and the extraction time is 45 min.

10. The method according to claim 1, characterized in that The gas chromatography-mass spectrometry analysis corresponding conditions include: Chromatographic conditions: HP-5MS capillary column; programmed temperature conditions: initial column temperature 40°C, hold for 1 min, increase to 80°C at 5°C / min, hold for 1 min, then increase to 260°C at 20°C / min, hold for 5 min; carrier gas is high-purity helium, constant flow mode, flow rate is 1.0 mL / min; splitless injection, solvent delay is 3 min; inlet temperature is 280°C; vacuum compensation and carrier gas saving are turned on, flow rate is 20.0 mL / min, hold for 2 min; mass spectrometry conditions: electron bombardment ion source, electron energy is 70 eV, transmission line temperature is 280°C, ion source temperature is 280°C; scanning mass range: 40-450 amu, scanning time: 0.2 s; filament current: 20 μA.

Images